Method for monitoring undisturbed sleep and related devices

Abstract

The application discloses a sleep disturbance-free monitoring method and related device, the method comprising: extracting features from a first piezoelectric signal to obtain body movement features, the first piezoelectric signal being obtained by superimposing a plurality of pre-processed original piezoelectric signals; performing frequency domain analysis on a second piezoelectric signal, identifying the frequency components corresponding to the wave peaks in the sleep heart rate correlation frequency band, extracting the frequency pairs corresponding to each time point to obtain a frequency set, the frequency pair comprising a plurality of frequency components in a first frequency sequence and the frequency of each frequency component, the first frequency sequence being obtained by arranging the frequency components in descending order according to the amplitude intensity; determining the heart rate features of each time point according to the frequency set; extracting features from the obtained piezoresistive signal to obtain piezoresistive features; and determining the sleep state according to the body movement features, the heart rate features of each time point, and the piezoresistive features. The application can improve the reliability and accuracy of sleep state recognition.

Description

Technical Field

[0001] This application relates to the field of sleep monitoring technology, and in particular to a sleep monitoring method and related device that does not disturb sleep. Background Technology

[0002] As living standards continue to improve, people are paying more and more attention to their health, including sleep health. This demand has directly driven the rapid development and widespread adoption of wearable devices such as smart bracelets and smartwatches. These devices, with their convenient operation and real-time data feedback, can provide users with basic monitoring data such as sleep duration and sleep cycles, becoming a common tool for many to manage their sleep health independently. However, this requires users to wear the device continuously throughout the entire sleep process.

[0003] Wearing these devices may interfere with sleep. For example, smartwatches, due to their large size and hard materials, can easily cause a feeling of constriction and pressure when worn. Even small smart bracelets may cause users persistent discomfort due to issues such as the strap material rubbing against the skin or improper adjustment of the fit. This discomfort is amplified during sleep at night, not only reducing the ease of falling asleep but also potentially causing sleep interruptions and a decrease in the proportion of deep sleep, thus affecting the user's original sleep quality to some extent.

[0004] Furthermore, for individuals with light sleep or high psychological sensitivity, the presence of wearable devices can trigger latent anxiety or tension, and even directly lead to insomnia symptoms such as difficulty falling asleep and fragmented sleep. In this unnatural sleep state, the data recorded by the device cannot objectively and truthfully reflect the user's actual sleep status, ultimately leading to biased sleep monitoring results and affecting the accuracy of the sleep monitoring findings. Summary of the Invention

[0005] This application provides a method and related apparatus for monitoring sleep without disturbance, so as to improve the reliability and accuracy of sleep state identification.

[0006] In a first aspect, embodiments of this application provide a method for monitoring sleep without disturbance, comprising:

[0007] Feature extraction is performed on the first piezoelectric signal to obtain body motion features. The first piezoelectric signal is obtained by superimposing multiple preprocessed original piezoelectric signals. A single original piezoelectric signal includes multiple time points.

[0008] Frequency domain analysis is performed on the second piezoelectric signal to identify the frequency components corresponding to the peaks within the sleep heart rate correlation frequency band. Frequency pairs corresponding to each time moment are extracted to obtain a frequency set. The frequency pairs include the preceding multiple frequency components in the first frequency sequence and the frequency of each frequency component. The first frequency sequence is obtained by arranging the frequency components in descending order according to their amplitude intensity. The second piezoelectric signal is the piezoelectric signal with the highest signal quality among the multiple original piezoelectric signals. The signal quality is determined based on a preset number of sampling values ​​in a single original piezoelectric signal.

[0009] The heart rate characteristics at each moment are determined based on the frequency set;

[0010] Feature extraction is performed on the acquired piezoresistive signal to obtain piezoresistive features;

[0011] Sleep state is determined based on the body movement characteristics, the heart rate characteristics at each moment, and the piezoresistive characteristics.

[0012] The step of extracting features from the first piezoelectric signal to obtain body motion features includes:

[0013] The peaks of the first piezoelectric signal are determined to obtain the peak set;

[0014] Determine the mean value of the peaks in the peak set;

[0015] Based on the peak set and the peak mean, a first number of abnormal peaks is determined, wherein the abnormal peaks are used to indicate peaks in the peak set that are greater than a preset multiple of the peak mean.

[0016] The body motion threshold is determined based on the median value of the first piezoelectric signal;

[0017] The body movement characteristics are determined based on the first quantity, the body movement threshold, and the first piezoelectric signal.

[0018] The step of determining the body movement characteristics based on the first quantity, the body movement threshold, and the first piezoelectric signal includes:

[0019] If the first quantity is detected to be non-zero, then the target body motion value corresponding to the piezoelectric signal at each moment in the first piezoelectric signal is determined according to the body motion threshold.

[0020] The body motion characteristics are determined based on the target body motion value;

[0021] If the first quantity is detected to be zero, then the initial body movement value is obtained;

[0022] The body movement characteristics are determined based on the initial body movement values.

[0023] The step of determining the heart rate characteristics at each moment based on the frequency set includes:

[0024] The heart rate value at each moment is determined based on each frequency pair in the frequency set;

[0025] Determine the signal quality of the piezoelectric signal at each moment in the second piezoelectric signal;

[0026] The heart rate characteristics are determined based on the heart rate value and the signal quality.

[0027] The step of determining the heart rate characteristics based on the heart rate value and the signal quality includes:

[0028] Get the first heart rate value at the current moment and the second heart rate value at the previous moment;

[0029] If the signal quality of the piezoelectric signal at the current moment is detected to be greater than or equal to a preset signal quality, then a first weight and a second weight are determined based on the signal quality of the piezoelectric signal at the current moment. The first weight is the weight of the first heart rate value, and the second weight is the weight of the second heart rate value.

[0030] The heart rate characteristics at the current moment are determined based on the first weight, the second weight, the first heart rate value, and the second heart rate value.

[0031] If the signal quality of the piezoelectric signal at the current moment is detected to be less than the preset signal quality, then the initial heart rate value is obtained;

[0032] The initial heart rate value is determined as the heart rate characteristic at the current moment.

[0033] The piezoresistive features include a first feature and a second feature. The step of extracting features from the acquired piezoresistive signal to obtain the piezoresistive features includes:

[0034] The first feature is obtained by determining the first number of times that the piezoresistive signal has a piezoresistive resistance less than a first preset threshold.

[0035] The second feature is obtained by determining the second number of times the piezoresistive signal is greater than the second preset threshold, wherein the second preset threshold is greater than the first preset threshold, and the second preset threshold and the first preset threshold are integer multiples of each other.

[0036] The step of determining the sleep state based on the body movement characteristics, the heart rate characteristics at each moment, and the pressure resistance characteristics includes:

[0037] Based on the described body movement characteristics, the sleep state is identified to obtain the first state;

[0038] Based on the described body movement characteristics or the first characteristic, a waking state is identified to obtain a second state;

[0039] Based on the body movement characteristics, the heart rate characteristics at each moment, the first characteristic, and the second characteristic, sleep stage identification is performed to obtain the third state;

[0040] The sleep state is obtained based on the first state, the second state, and the third state.

[0041] Secondly, embodiments of this application provide a sleep monitoring device that does not disturb sleep, comprising:

[0042] The first extraction unit is used to extract features from the first piezoelectric signal to obtain body motion features. The first piezoelectric signal is obtained by superimposing multiple pre-processed original piezoelectric signals. A single original piezoelectric signal includes multiple time points.

[0043] The analysis unit is used to perform frequency domain analysis on the second piezoelectric signal, identify the frequency components corresponding to the peaks within the sleep heart rate correlation frequency band, extract the frequency pairs corresponding to each time moment, and obtain a frequency set. The frequency pairs include the preceding multiple frequency components in the first frequency sequence and the frequency of each frequency component. The first frequency sequence is obtained by arranging the frequency components in descending order according to their amplitude intensity. The second piezoelectric signal is the piezoelectric signal with the highest signal quality among the multiple original piezoelectric signals. The signal quality is determined based on a preset number of sampling values ​​in a single original piezoelectric signal.

[0044] The first determining unit is used to determine the heart rate characteristics at each moment based on the frequency set;

[0045] The second extraction unit is used to extract features from the acquired piezoresistive signal to obtain piezoresistive features;

[0046] The second determining unit is used to determine the sleep state based on the body movement characteristics, the heart rate characteristics at each moment, and the piezoresistive characteristics.

[0047] Thirdly, embodiments of this application provide an electronic device, including a memory, a processor, and executable program code stored in the memory and executable on the processor, wherein the processor executes the executable program code and performs the steps of the method described in the first aspect.

[0048] Fourthly, embodiments of this application provide a computer-readable storage medium storing executable program code, the executable program code including execution instructions for performing the steps of the method as described in the first aspect.

[0049] Fifthly, embodiments of this application provide a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps described in the first aspect of embodiments of this application. The computer program product may be a software installation package.

[0050] As can be seen, in this embodiment, feature extraction is first performed on the first piezoelectric signal to obtain body motion features. The first piezoelectric signal is obtained by superimposing multiple preprocessed original piezoelectric signals, and a single original piezoelectric signal includes multiple time points. Then, frequency domain analysis is performed on the second piezoelectric signal to identify the frequency components corresponding to the peaks within the sleep heart rate correlation frequency band. Frequency pairs corresponding to each time point are extracted to obtain a frequency set. The frequency pairs include the preceding multiple frequency components in the first frequency sequence and the frequency of each frequency component. The first frequency sequence is obtained by arranging the frequency components in descending order according to their amplitude intensity. The second piezoelectric signal is the piezoelectric signal with the highest signal quality among the multiple original piezoelectric signals. The signal quality is determined based on a preset number of sampling values ​​in a single original piezoelectric signal. Then, the heart rate features at each time point are determined based on the frequency set. Next, feature extraction is performed on the acquired piezoresistive signal to obtain piezoresistive features. Finally, the sleep state is determined based on the body motion features, the heart rate features at each time point, and the piezoresistive features.

[0051] This application extracts body motion features from a first piezoelectric signal obtained by superimposing multiple original piezoelectric signals, which reduces single-channel errors and makes body motion judgment more stable. Furthermore, it selects a second piezoelectric signal with the best signal quality for frequency domain analysis, screening key frequency components to extract heart rate features, ensuring the accuracy of heart rate feature extraction. Through the above differentiated processing of multiple piezoelectric signals, the accuracy of feature extraction can be improved. Combined with the piezoresistive features extracted from the piezoresistive signal, a three-dimensional feature system of body motion, heart rate, and piezoresistive signals is formed, collaboratively judging sleep state, making sleep monitoring more comprehensive, and improving the reliability and accuracy of sleep state recognition. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 This application provides a system architecture diagram of a sleep monitoring system.

[0054] Figure 2This is a schematic flowchart of a sleep monitoring method that does not disturb sleep, provided in an embodiment of this application.

[0055] Figure 3 This is a schematic diagram of a process for extracting body motion features provided in an embodiment of this application;

[0056] Figure 4 This is a schematic diagram of a process for extracting heart rate features provided in an embodiment of this application;

[0057] Figure 5 This is a flowchart illustrating another sleep monitoring method that does not disturb sleep, provided in an embodiment of this application.

[0058] Figure 6 This is a block diagram of the functional units of a sleep monitoring device that does not disturb sleep, provided in an embodiment of this application.

[0059] Figure 7 This is a block diagram of the functional units of another sleep monitoring device provided in this application embodiment;

[0060] Figure 8 This is a schematic diagram of the structure of an electronic device proposed in an embodiment of this application. Detailed Implementation

[0061] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0062] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0063] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0064] As living standards continue to improve, people are paying more and more attention to their health, including sleep health. This demand has directly driven the rapid development and widespread adoption of wearable devices such as smart bracelets and smartwatches. These devices, with their convenient operation and real-time data feedback, can provide users with basic monitoring data such as sleep duration and sleep cycles, becoming a common tool for many to manage their sleep health independently. However, this requires users to wear the device continuously throughout the entire sleep process.

[0065] Wearing these devices may interfere with sleep. For example, smartwatches, due to their large size and hard materials, can easily cause a feeling of constriction and pressure when worn. Even small smart bracelets may cause users persistent discomfort due to issues such as the strap material rubbing against the skin or improper adjustment of the fit. This discomfort is amplified during sleep at night, not only reducing the ease of falling asleep but also potentially causing sleep interruptions and a decrease in the proportion of deep sleep, thus affecting the user's original sleep quality to some extent.

[0066] Furthermore, wearable devices have limited battery life, and the detected body movement data is easily affected by the user's sleeping posture and the device's position. Additionally, for individuals with light sleep or high psychological sensitivity, the presence of wearable devices can trigger latent anxiety or tension, and even directly lead to insomnia symptoms such as difficulty falling asleep and fragmented sleep. In this unnatural sleep state, the data recorded by the device cannot objectively and truthfully reflect the user's actual sleep status, ultimately leading to biased sleep monitoring results and affecting their accuracy.

[0067] Among these, sleep monitoring can be conducted using non-contact monitoring devices such as radar and infrared sensors. These devices detect the user's breathing, heart rate, and body movement data by emitting and receiving microwave, ultrasonic, or infrared signals. While these devices do not require direct contact with the user's body, they are not accurate enough in detecting subtle physiological signals and are susceptible to environmental factors such as temperature, humidity, and electromagnetic interference. Some devices also require cameras or infrared sensors, which may raise privacy concerns for users.

[0068] To address the aforementioned problems, this application provides a method and related apparatus for monitoring sleep without disturbance. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0069] Please see Figure 1 , Figure 1 This is a system architecture diagram of a sleep monitoring system provided in an embodiment of this application. Figure 1 As shown, the sleep monitoring system 100 includes a data acquisition module 101, a printed circuit board module 102, a signal amplification and filtering module 103, a data preprocessing module 104, a feature extraction module 105, a sleep state recognition module 106, a wireless transmission module 107, and a user interface module 108, wherein the above modules are interconnected and communicate with each other.

[0070] The data acquisition module 101 includes multiple piezoelectric sensors and piezoresistive sensors. The data acquisition of the multiple piezoelectric sensors is independent and does not affect each other. The piezoelectric sensors are used to collect minute vibration signals such as the user's breathing and heartbeat. The piezoresistive sensors are used to detect changes in bed surface pressure to obtain information about the user's body position and turning behavior. They capture the physical characteristics of human movement through changes in the resistance of pressure-sensitive materials, reflecting the user's sleeping posture and movement.

[0071] The printed circuit board module 102 is responsible for the circuit control, signal transmission, power management and signal interface of the sleep monitoring system 100. It is communicatively connected to the data acquisition module 101 and the signal amplification and filtering module 103, and provides a stable current and data transmission path for the entire system.

[0072] The signal amplification and filtering module 103 is used to amplify and filter the signal acquired by the data acquisition module 101 to ensure the quality of the sensor signal and make it more suitable for subsequent data acquisition and analysis.

[0073] Specifically, the minute vibration signals from the piezoelectric sensor are amplified to ensure that the signal strength meets the requirements for subsequent data acquisition.

[0074] Among them, a finite impulse response (FIR) low-pass filter can be used to eliminate 50Hz power frequency noise and other high-frequency interference in piezoelectric signals.

[0075] The data preprocessing module 104 is used to convert the amplified and filtered sensor signals into digital data and preprocess them. For example, it performs differential processing, downsampling processing, and signal superposition processing on multiple piezoelectric signals.

[0076] The feature extraction module 105 is used to extract body movement, heart rate and piezoresistive features from piezoelectric signals and piezoresistive signals to facilitate subsequent sleep state analysis.

[0077] The sleep state recognition module 106 is used to identify sleep onset, wakefulness, and sleep stages based on body movement characteristics, heart rate characteristics, and piezoresistive characteristics, thereby obtaining the sleep state.

[0078] The wireless transmission module 107 is used to transmit the processed sleep data to the user's mobile device via Wi-Fi or Bluetooth, so that the user can view the sleep report in real time.

[0079] The user interface module 108 provides users with sleep quality reports, duration distribution of each sleep stage, and body movement and heart rate change curves via a mobile application. It supports historical data visualization to help users track their sleep status over the long term.

[0080] Based on this, this application provides a method for monitoring sleep without disturbance, and the following is a detailed description of this application with reference to the accompanying drawings.

[0081] Please see Figure 2 , Figure 2 This is a flowchart illustrating a sleep monitoring method without disturbance provided in an embodiment of this application. Figure 2 As shown, the method includes the following steps:

[0082] S210, feature extraction is performed on the first piezoelectric signal to obtain the body motion features.

[0083] The first piezoelectric signal is obtained by superimposing multiple preprocessed original piezoelectric signals, and a single original piezoelectric signal includes multiple moments.

[0084] The system collects multiple piezoelectric signals through multiple piezoelectric sensors and piezoresistive signals through a piezoresistive sensor. These piezoelectric and piezoresistive sensors can be placed under the sheets or blankets to detect sleep-related physiological parameters and provide an undisturbed sleep monitoring experience.

[0085] Piezoelectric signals are easily affected by noise, and the data may contain 50Hz power frequency signals. Therefore, it is necessary to filter the multiple piezoelectric signals collected by multiple piezoelectric sensors to remove power frequency signals and noise.

[0086] The frequencies of respiration and heart rate are within 3 Hz, and the frequency of body movement is within 10 Hz. In order to effectively eliminate power frequency signals and noise, the piezoelectric signals were processed by an FIR low-pass filter.

[0087] For example, the FIR low-pass filter employs a least-squares design method, with both the passband weight (Wpass) and stopband weight (Wstop) set to 1. This ensures that the flatness of the useful signal in the passband and the attenuation effect of the interference signal in the stopband are given equal importance. The filter order is set to 20, effectively controlling the computational load while ensuring filtering performance, thus adapting to the needs of real-time signal processing. Its sampling frequency is 200Hz, the passband cutoff frequency (Fpass) is 10Hz, allowing low-frequency useful signals such as body movement and breathing to pass through without significant attenuation, and the stopband cutoff frequency (Fstop) is 35Hz, which can significantly suppress 50Hz power frequency noise and other high-frequency interference signals, ultimately filtering out redundant interference.

[0088] In one possible embodiment, multiple independent piezoelectric sensors are deployed to synchronously and continuously acquire piezoelectric signals, resulting in multiple long sequences of piezoelectric signals. Then, for each long sequence of piezoelectric signals, several data segments are continuously extracted from the long sequence according to preset time window parameters. Each extracted data segment is the original piezoelectric signal for one time window, and each original piezoelectric signal includes sampled data from multiple consecutive time points.

[0089] Each piezoelectric sensor has the same time window. For example, the time window is one minute.

[0090] In this process, the long sequence of piezoelectric signals from each piezoelectric sensor is broken down into multiple raw piezoelectric signals for different time windows. Based on the synchronized time windows, the raw piezoelectric signals from the same time window corresponding to different piezoelectric sensors are integrated to obtain multiple raw piezoelectric signals corresponding to each synchronized time window.

[0091] Each original piezoelectric signal undergoes preprocessing, followed by signal superposition to obtain the first piezoelectric signal. The preprocessing includes differential processing and downsampling.

[0092] Specifically, the user's body movement signal is a high-frequency signal. In order to obtain the high-frequency signal and reduce the influence of low-frequency noise, differential processing is performed on each piezoelectric signal to amplify the abrupt changes in the signal. The calculation formula for the differential processing is as follows:

[0093]

[0094] in, This represents the piezoelectric signal after differential processing. This represents the piezoelectric data at time n. This represents the piezoelectric data at time n-1, where N represents the length of the piezoelectric signal.

[0095] To reduce the computational load of the algorithm without affecting the results, the piezoelectric data after differential processing is downsampled. For example, the signal sampling frequency is reduced from 200Hz to 20Hz. The downsampling calculation formula is as follows:

[0096]

[0097] in, This is the piezoelectric signal after downsampling.

[0098] During body movement, multiple piezoelectric signals fluctuate significantly. Superimposing these signals amplifies the fluctuations. At other times, the signals are noise signals. Since noise signals are random, their superposition does not amplify them. Thus, multiple piezoelectric signals are superimposed into a single piezoelectric signal to enhance the piezoelectric changes caused by body movement, thereby amplifying the body movement signal.

[0099] Since the focus is on the piezoelectric changes caused by volumetric motion rather than the specific piezoelectric value at a particular moment, the absolute value of the piezoelectric data after differential processing needs to be taken during the superposition process. The specific calculation formula is as follows:

[0100]

[0101] in, This represents the piezoelectric signal after superposition processing, i.e., the first piezoelectric signal. This is the first piezoelectric signal. This is the second piezoelectric signal. For the i-th piezoelectric signal, Indicates the length of the signal. This indicates taking the absolute value of the signal.

[0102] In one possible embodiment, please refer to Figure 3 , Figure 3 This is a schematic diagram of a process for extracting body motion features provided in an embodiment of this application. Figure 3 The flowchart illustrates the steps for extracting features from the first piezoelectric signal to obtain body motion features. The specific steps are as follows:

[0103] S310, determine the peak of the first piezoelectric signal to obtain the peak set.

[0104] The criteria for determining the peak include: the piezoelectric data at time n is greater than the piezoelectric data at the previous time; the piezoelectric data at time n is greater than the piezoelectric data at the next time; the piezoelectric data at time n is greater than a minimum threshold; and the piezoelectric data at time n is less than a maximum threshold. For example, the minimum threshold can be 8000, and the maximum threshold can be 500000.

[0105] Among them, the peaks in the first piezoelectric signal are identified according to the above-mentioned judgment conditions. This yields the peak set.

[0106] S320, determine the mean value of the peaks in the peak set.

[0107] After obtaining the peak set, the peak mean is calculated. The calculation formula is as follows:

[0108]

[0109] in, This represents the number of peaks in the peak cluster.

[0110] S330, determine the first number of abnormal peaks based on the peak set and the peak mean.

[0111] The abnormal peaks are used to indicate peaks that are greater than a preset multiple of the average peak value.

[0112] Among them, abnormal peaks are peaks with exceptionally large values ​​compared to other peaks, corresponding to the user's body movements. A peak is identified as abnormal if it exceeds a preset multiple of the average peak value. Preferably, the preset multiple is 2.5.

[0113] Specifically, the first number of abnormal peaks The calculation formula is as follows:

[0114]

[0115] Where I(x) represents an exponential function, if I(x) is true, i.e. Greater than I(x) = 1 if x is positive, otherwise it is 0.

[0116] S340, determine the body motion threshold based on the median of the first piezoelectric signal.

[0117] The magnitude of the piezoelectric signal varies depending on the body motion, the point of force application, and the placement of the piezoelectric sensor. Therefore, when calculating the body motion, the body motion threshold needs to be dynamically adjusted based on the median of the time window.

[0118] Among them, the first piezoelectric signal Sort in ascending order to get The formula for calculating the median m is as follows:

[0119]

[0120] Where m is the median of e(n) within the time window, for The results after sorting in ascending order.

[0121] In one possible embodiment, an initial body movement threshold is obtained. If the initial body movement threshold is less than or equal to a preset minimum value (low), the body movement threshold is adjusted to the preset minimum value (low). If the initial body movement threshold is within the range of the preset minimum value (low) and the preset maximum value (high), the body movement threshold is adjusted to the median of a preset multiple. Preferably, the body movement threshold can be adjusted to 9m. If the initial body movement threshold is greater than or equal to the range of the preset maximum value (high), the body movement threshold is adjusted to the preset maximum value (high).

[0122] For example, the value of low can be 2e5, and the value of high can be 5e5. (Body movement threshold) The specific calculation formula is as follows:

[0123]

[0124] S350, determine the body movement characteristics based on the first quantity, the body movement threshold, and the first piezoelectric signal.

[0125] Among them, the calculation method for determining the body movement characteristics based on the first quantity.

[0126] In one possible embodiment, determining the body movement feature based on the first quantity, the body movement threshold, and the first piezoelectric signal includes: if the first quantity is detected to be non-zero, then determining the target body movement value corresponding to the piezoelectric signal at each moment in the first piezoelectric signal based on the body movement threshold; determining the body movement feature based on the target body movement value; if the first quantity is detected to be zero, then obtaining an initial body movement value; and determining the body movement feature based on the initial body movement value.

[0127] Specifically, if the number of abnormal peaks in the first piezoelectric signal corresponding to the current time window is zero, an initial body motion value is obtained. This initial body motion value is used to characterize the absence of body motion events, and the body motion feature is zero. If the number of abnormal peaks in the first piezoelectric signal corresponding to the current time window is not zero, each piezoelectric data point in the first piezoelectric signal is compared with a body motion threshold. If the piezoelectric data point is greater than or equal to the body motion threshold, the target body motion value is one; if the piezoelectric data point is less than the body motion threshold, the target body motion value is zero. The sum of the target body motion values ​​is calculated to obtain the body motion feature.

[0128] Among them, the target body dynamic value The calculation formula is as follows:

[0129]

[0130] in, Let I(x) be the first quantity, and let I(x) be an exponential function. If I(x) is true, then... Greater than or equal to I(x) = 1 if x is positive, otherwise it is 0.

[0131] As can be seen, in this embodiment, the presence of body movement is reflected by the number of abnormal peaks. Combined with a body movement threshold determined based on the characteristics of the first piezoelectric signal, accurate screening of body movement-related signals can be achieved, effectively eliminating irrelevant noise interference. Simultaneously, by combining the raw data of the first piezoelectric signal, body movement events can be quickly identified through the number of abnormal peaks, and body movement thresholds can be used to further distinguish body movement responses of different intensities. Feature extraction through multi-dimensional information fusion demonstrates the accuracy of the judgment and the completeness of the data, ensuring that body movement features truly reflect the actual body movement state and providing reliable data support for subsequent sleep monitoring.

[0132] S220 performs frequency domain analysis on the second piezoelectric signal, identifies the frequency components corresponding to the peaks within the sleep heart rate correlation frequency band, extracts the frequency pairs corresponding to each time moment, and obtains the frequency set.

[0133] The frequency pair includes a plurality of preceding frequency components in a first frequency sequence and the frequency of each frequency component. The first frequency sequence is obtained by arranging the frequency components in descending order according to their amplitude intensity. The second piezoelectric signal is the piezoelectric signal with the highest signal quality among the plurality of original piezoelectric signals. The signal quality is determined based on a preset number of sampling values ​​in a single original piezoelectric signal.

[0134] Among them, the piezoelectric signal with the best signal quality is dynamically selected from multiple original piezoelectric signals and used as the second piezoelectric signal.

[0135] In one possible embodiment, two piezoelectric sensors are deployed to acquire two piezoelectric signals x1(n) and x2(n). The two piezoelectric signals are then downsampled, for example, by reducing the sampling frequency from 200Hz to 20Hz. The sum of the absolute values ​​of all samples included at each moment in each piezoelectric signal is then calculated. The specific formula is as follows:

[0136]

[0137] in, The number of samples at each time moment is used to represent the preset number of sample values ​​included at each time moment, for example, It is 20. This is the piezoelectric signal after downsampling.

[0138] Specifically, according to the above formula, the sum of absolute values ​​of x1(n) at each moment, s1, and the sum of absolute values ​​of x2(n) at each moment, s2, are calculated. The sums of s1 at multiple moments are obtained as S1; the sums of s2 at multiple moments are obtained as S2. The ratio of S2 to S1 is calculated. If this ratio is greater than a preset threshold, x1(n) is determined as the second piezoelectric signal; otherwise, x2(n) is determined as the second piezoelectric signal. Preferably, the number of moments for which s1 is summed is 6, and the preset threshold is 1.2.

[0139] In one possible embodiment, if the number of piezoelectric sensors is greater than two, then for each acquired piezoelectric signal, the same preset number of sample values ​​are extracted, denoised, and baseline corrected. Then, based on the preset number of sample values, the signal-to-noise ratio is calculated, the coefficient of variation of the preset number of sample values ​​is calculated, and the similarity to the standard piezoelectric signal is calculated. The above data are normalized, weighted according to importance, and a comprehensive score is calculated. The piezoelectric signal with the highest score is the one with the best signal quality, i.e., the second piezoelectric signal. The coefficient of variation is used to indicate the standard deviation or mean of the preset number of sample values, etc.

[0140] Among them, the second piezoelectric signal Perform a fast Fourier transform to obtain , Indicates the index of the frequency component. It is 256. It is the imaginary unit, and its formula is as follows:

[0141]

[0142] Among them, the sleep heart rate associated frequency band refers to the specific frequency range corresponding to the human heart rate during sleep, specifically 0.66-3.33Hz, which covers the normal physiological heart rate range during human sleep.

[0143] Among them, in the transformed frequency domain signal The study focuses on the sleep heart rate correlation band of 0.66-3.33 Hz, identifying all peaks within this band. Each peak corresponds to a frequency component with significant energy, and these peaks are arranged in descending order of amplitude intensity to obtain the first frequency sequence. The higher the amplitude intensity, the more likely the frequency component is to be a valid heart rate-related signal.

[0144] Each time point corresponds to a first frequency sequence.

[0145] Specifically, several preceding frequency components and the specific frequency value of each selected preceding frequency component are selected from each first frequency sequence to obtain several frequency pairs selected at each time point, thereby obtaining a frequency set.

[0146] Preferably, the first five frequency components and the specific frequency values ​​of each selected preceding frequency component can be selected from the first frequency sequence to obtain five frequency pairs, and a frequency set can be constructed based on the five frequency pairs selected at each time.

[0147] S230, determine the heart rate characteristics at each moment based on the frequency set.

[0148] In one possible embodiment, please refer to Figure 4 , Figure 4 This is a schematic diagram of a process for extracting heart rate features provided in an embodiment of this application. Figure 4 The flowchart illustrates the steps for determining the heart rate characteristics at each moment based on a set of frequencies. The specific steps are as follows:

[0149] S410, determine the heart rate value at each moment based on each frequency pair in the frequency set.

[0150] The frequency set includes several frequency pairs selected at each time moment. Based on the several frequency pairs selected at each time moment, the heart rate value at each time moment is determined.

[0151] Specifically, the frequency domain of a heart rate signal not only includes the fundamental frequency representing the heart rate itself, but also contains harmonics that are integer multiples of the fundamental frequency. That is, if the fundamental frequency is f, it will be accompanied by harmonics such as 2f and 3f. Based on the harmonic characteristics of the heart rate signal, the frequencies corresponding to the first few peaks in the first frequency sequence are screened. Frequency combinations with integer multiple relationships are searched among several candidate frequencies. Typically, the fundamental frequency is the lowest frequency among these, and the other higher frequencies must be integer multiples of the lower frequency, such as 2 or 3. Through this integer multiple correlation, the peak most likely to be the heart rate fundamental frequency is determined from several candidate peaks. Then, based on the frequency value corresponding to the peak most likely to be the heart rate fundamental frequency, the heart rate value at that moment is calculated using the formula: "Heart rate = frequency × 60".

[0152] S420, determine the signal quality of the piezoelectric signal at each moment in the second piezoelectric signal.

[0153] Specifically, for each moment Calculate the total energy within the 0.66-3.33Hz frequency spectrum. The calculation formula is as follows:

[0154]

[0155] in, The k-index corresponding to 0.66Hz. The k subscript corresponds to 3.33Hz.

[0156] Among them, noise energy is calculated above 3.33Hz. The calculation formula is as follows:

[0157]

[0158] in, The k-index corresponds to 10Hz.

[0159] Among them, based on total energy and noise energy Calculate the signal-to-noise ratio of the signal. The specific formula is as follows:

[0160]

[0161] Among them, according to the signal-to-noise ratio Determine signal quality signal quality The value can be -1, 0, or 1. -1 indicates very poor signal quality, 1 indicates good signal quality, and 0 indicates moderate signal quality. The specific calculation formula is as follows:

[0162]

[0163] The above steps for calculating signal quality are repeated to obtain the signal quality at each time point.

[0164] S430, determine the heart rate characteristics based on the heart rate value and the signal quality.

[0165] Specifically, the weights of the heart rate values ​​at the previous and current times are dynamically assigned based on the signal quality at each time point, and a weighted sum is performed to determine the heart rate characteristics at the current time point, thereby obtaining the heart rate characteristics at each time point.

[0166] In one possible embodiment, determining the heart rate feature based on the heart rate value and the signal quality includes: acquiring a first heart rate value at the current moment and a second heart rate value at the previous moment; detecting that the signal quality of the piezoelectric signal at the current moment is greater than or equal to a preset signal quality, then determining a first weight and a second weight based on the signal quality of the piezoelectric signal at the current moment, wherein the first weight is the weight of the first heart rate value and the second weight is the weight of the second heart rate value; determining the heart rate feature at the current moment based on the first weight, the second weight, the first heart rate value, and the second heart rate value; detecting that the signal quality of the piezoelectric signal at the current moment is less than the preset signal quality, then acquiring an initial heart rate value; and determining the initial heart rate value as the heart rate feature at the current moment.

[0167] Specifically, when the signal quality is less than a preset signal quality, the initial heart rate value is determined as the heart rate characteristic at the current moment. When the signal quality is greater than or equal to the preset signal quality, the weights of the first and second heart rate values ​​are determined based on the signal quality at the current moment. These weights are then combined and summed to obtain the heart rate characteristic at the current moment. This process is repeated to obtain the heart rate characteristic at each moment.

[0168] In one possible embodiment, the preset signal quality can be 0. If the calculated signal quality at the current moment is -1, indicating very poor signal quality, then an initial heart rate value of 255 is obtained and used as the heart rate feature at the current moment. If the calculated signal quality at the current moment is 0, equal to the preset signal quality, indicating moderate signal quality, a weight of 0.25 can be assigned to the first heart rate value and a weight of 0.75 to the second heart rate value, and these are combined and weighted to obtain the heart rate feature at the current moment. If the calculated signal quality at the current moment is 1, greater than the preset signal quality, indicating good signal quality, a weight of 0.125 can be assigned to the first heart rate value and a weight of 0.875 to the second heart rate value, and these are combined and weighted to obtain the heart rate feature at the current moment. The specific formula is as follows:

[0169]

[0170] in, The first heart rate value, This is the second heart rate value. This represents the heart rate characteristics at the current moment.

[0171] In one possible embodiment, the overall signal quality of the second piezoelectric signal can be calculated, and the weights of the first heart rate value and the second heart rate value can be determined based on the overall signal quality. That is, the weights of the first heart rate value corresponding to each moment in the second piezoelectric signal are equal, and the weights of the second heart rate value corresponding to each moment are equal.

[0172] As can be seen, in this embodiment, heart rate values ​​at each moment are accurately extracted, and signal quality is simultaneously evaluated. When the signal quality meets the standard, the weights of the current and previous heart rate values ​​are dynamically allocated, reflecting both the timeliness of real-time data and smoothing fluctuations with historical data to avoid the impact of single-point anomalies. When the standard is not met, the initial heart rate value is used to avoid feature distortion caused by poor signals. This embodiment improves the accuracy of heart rate features by filtering noise interference and ensures data continuity and stability through temporal weight balancing, while avoiding feature loss problems when signal quality is poor, providing reliable data support for subsequent sleep monitoring.

[0173] S240 performs feature extraction on the acquired piezoresistive signal to obtain piezoresistive features.

[0174] In one possible embodiment, the piezoresistive feature includes a first feature and a second feature. The step of extracting features from the acquired piezoresistive signal to obtain the piezoresistive feature includes: determining a first number of times the piezoresistive signal has a piezoresistive resistance less than a first preset threshold to obtain the first feature; determining a second number of times the piezoresistive signal has a piezoresistive resistance greater than a second preset threshold to obtain the second feature, wherein the second preset threshold is greater than the first preset threshold, and the second preset threshold and the first preset threshold are integer multiples of each other.

[0175] The time window of the piezoresistive signal is the same as that of the second piezoelectric signal, preferably 1 minute.

[0176] Preferably, the first preset threshold is 100 and the second preset threshold is 1500, that is, the number of times the piezoresistive signal with a piezoresistive resistance less than 100 is determined as C1, which is the first feature, and the number of times the piezoresistive signal with a piezoresistive resistance greater than 1500 is determined as C2, which is the second feature.

[0177] S250, determine the sleep state based on the body movement characteristics, the heart rate characteristics at each moment, and the piezoresistive characteristics.

[0178] In one possible embodiment, determining the sleep state based on the body movement characteristics, the heart rate characteristics at each moment, and the piezoresistive characteristics includes: identifying a sleep state based on the body movement characteristics to obtain a first state; identifying a wakefulness state based on the body movement characteristics or the first characteristic to obtain a second state; identifying sleep stages based on the body movement characteristics, the heart rate characteristics at each moment, the first characteristic, and the second characteristic to obtain a third state; and obtaining the sleep state based on the first state, the second state, and the third state.

[0179] The rules for recognizing sleep state and wakefulness can be obtained based on expert experience and analysis of collected sleep data, including but not limited to those mentioned in the following embodiments.

[0180] In one possible implementation, when the body movement characteristic b value is less than 10, it indicates that the user has hardly engaged in any physical activity. If this state continues for 10 minutes, it is determined that the user is asleep.

[0181] In one possible embodiment, if short-term high-intensity activity is determined based on body movement characteristics, the user is considered awake. For example, if the body movement characteristic b calculated every minute is greater than 10 within any consecutive 4 minutes, and the total body movement characteristic value b exceeds 400, the user is considered awake.

[0182] In one possible embodiment, if it is determined that there is a medium-to-long duration of body movement based on body movement characteristics, then it is determined that the user is awake. For example, if the body movement characteristic b calculated every minute is greater than 10 within any consecutive 10 minutes, and the total body movement characteristic value b exceeds 400 during this period, then it is determined that the user is awake.

[0183] In one possible embodiment, whether the user is awake is determined based on the piezoresistive characteristics, that is, if the piezoresistive characteristic C1 is greater than 120 within any time window, then the user is judged to be awake.

[0184] In one possible embodiment, sleep staging can be based on a machine learning model, with input features including body movement features, heart rate features, and piezoelectricity characteristics, and outputting classification labels for wakefulness, light sleep, deep sleep, and REM sleep. For example, the model can be trained based on 50 nights of sleep data, i.e., 20,000 samples.

[0185] The data acquisition included piezoelectric and piezoresistive signals from the sleep monitoring band, as well as polysomnography data from the user's sleep. The polysomnography data was used to label the piezoelectric and piezoresistive signals, providing accurate sleep stage labels for the training samples.

[0186] Each sample can be constructed with a 1-minute time window. First, the body motion feature 'b' is calculated per minute, and the heart rate is extracted per second, resulting in 60 heart rate features per minute. Then, the first feature C1 and the second feature C2 are calculated per minute. A training sample is then constructed based on one body motion feature, 60 heart rate features, and two piezoresistive features C1 and C2. The corresponding ground truth labels are assigned by aligning with polysomnography data. Through this process, data collection and training sample construction are completed.

[0187] Since the value ranges of different features may be different, the body motion feature, heart rate feature, and piezoresistive feature are normalized to improve the model performance.

[0188] In one possible embodiment, XGBoost performs well in handling noisy and nonlinear data, has strong generalization ability, and converges quickly when the sample size is moderate. Therefore, in this embodiment, the XGBoost model can be used for sleep staging. Since XGBoost can only process numerical labels, the classification labels are converted into numerical values. For example, wakefulness is 0, light sleep is 1, deep sleep is 2, and REM sleep is 3.

[0189] The main parameters of the XGBoost model include: maximum tree depth, used to control model complexity; learning rate, used to control the impact of each tree on the final model; number of trees, i.e., the number of model iterations; sampling ratio of each tree, used to prevent overfitting; proportion of features randomly sampled from each tree, used to control feature selection; gamma, used to control the minimum loss reduction for each split, reducing overfitting; setting the learning objective; and selecting the evaluation metric.

[0190] Preferably, the maximum tree depth is 6; the learning rate is 0.1; the number of trees is 300; the sampling ratio of each tree is 0.8; the feature ratio of random sampling in each tree is 0.8; the gamma is 0.1; the model is specified for multi-class classification tasks, the output result is the directly corresponding class label, and the number of classes is 4; the evaluation metric is multi-class log loss.

[0191] The process involves inputting sample features and their corresponding classification labels into the XGBoost model, training the model through supervised learning, and evaluating its performance using cross-validation. The trained XGBoost model is then used to predict new body movement, heart rate, and pressure resistance features, enabling real-time sleep staging.

[0192] By combining the identified sleep state, wakefulness state, and sleep stage status, a complete sleep structure profile and comprehensive sleep quality assessment can be obtained. This includes key parameters for the entire sleep cycle, such as total sleep time, sleep latency, number and duration of nighttime awakenings, and the proportion and duration of each sleep stage. It also includes the integrity and continuity of the sleep cycle, thereby accurately determining sleep quality, such as whether the proportion of deep sleep is sufficient, whether there is sleep fragmentation, difficulty falling asleep, or early awakening. Furthermore, based on this information, sleep disorders can be identified, and personalized sleep improvement plans can be developed.

[0193] As can be seen, in this embodiment, by extracting body motion features from the first piezoelectric signal obtained by superimposing multiple original piezoelectric signals, single-channel errors can be reduced, making body motion judgment more stable; and by selecting the second piezoelectric signal with the best signal quality for frequency domain analysis and screening key frequency components to extract heart rate features, the accuracy of heart rate features can be ensured. Through the above differential processing of multiple piezoelectric signals, the accuracy of feature extraction can be improved. Combined with the piezoresistive features extracted from the piezoresistive signal, a three-dimensional feature system of body motion, heart rate, and piezoresistive signals is formed, which collaboratively judges sleep state, making sleep monitoring more comprehensive and improving the reliability and accuracy of sleep state recognition.

[0194] In one possible embodiment, please refer to Figure 5 , Figure 5 This is a flowchart illustrating another undisturbed sleep monitoring method provided in this application embodiment, as shown below. Figure 5As shown, two piezoelectric sensors and one piezoresistive sensor are deployed under the user's mattress or sheet to collect two piezoelectric signals and one piezoresistive signal. The two piezoelectric signals are preprocessed separately, such as filtering power frequency interference, differential processing, and downsampling. Then, they are superimposed with the processed piezoelectric signals to obtain the first piezoelectric signal. Body movement features are extracted based on the first piezoelectric signal. The sleep state and wakefulness state are then determined based on the body movement features to obtain the first result.

[0195] Specifically, piezoresistive features are extracted based on the piezoresistive signal; the conscious state is determined based on the piezoresistive features, resulting in a second result.

[0196] Specifically, the piezoelectric signal with better signal quality from the two piezoelectric signals is selected to obtain the second piezoelectric signal; the spectrum of the second piezoelectric signal is analyzed and the heart rate is calculated to obtain the heart rate characteristics; the heart rate characteristics, body motion characteristics, and piezoresistive characteristics are normalized; the trained XGBoost model is input and the sleep stage status is output to obtain the third result.

[0197] The complete sleep state is obtained by combining the first, second, and third results.

[0198] As can be seen, in this embodiment, the non-contact design avoids the discomfort associated with wearable devices and eliminates the need for complex installation procedures, greatly improving ease of use. Simultaneously, the combination of piezoelectric and piezoresistive sensors provides multiple signal dimensions for sleep monitoring, significantly improving the reliability and accuracy of the monitoring signals. Based on the rich data collected from multiple sensors and the sleep staging algorithm, the accuracy of sleep state determination is further optimized.

[0199] In addition, the under-bed location and sensor selection reduce interference from environmental factors and avoid the problem of wearable devices being easily affected by movement and position changes, achieving a comfortable, convenient, anti-interference, and highly accurate sleep monitoring effect.

[0200] In one possible embodiment, this application can be applied to home sleep monitoring and health management, with specific application areas including smart homes, personal health management, and family health monitoring. Specifically, it can be used as a home smart health device, placed under the user's mattress or bed sheet to collect the user's sleep data in real time, and provide detailed sleep reports through a mobile application or smart home platform to help users understand their sleep quality, sleep stages, and abnormal sleep patterns.

[0201] In one possible embodiment, this application can also be applied to sleep research in hospitals and medical institutions. Specific application areas may include medical research, sleep disorder diagnosis and treatment, and rehabilitation care. Specifically, it can be used as an auxiliary tool for sleep research and sleep disorder diagnosis. Medical personnel can use this device to perform non-invasive sleep monitoring, record the patient's sleep stages and respiratory and heart rate characteristics in a natural sleep environment, and assist in the diagnosis of insomnia, sleep apnea, and other problems.

[0202] In one possible embodiment, this application can also be applied to nursing homes and elderly health management, with specific application areas including elderly care services, elderly health management, and long-term care. Specifically, the device can be used in nursing homes or elderly care centers to monitor the sleep quality of the elderly. By collecting sleep data in real time, caregivers can promptly detect abnormal sleep patterns or abnormal heart rate and respiratory rate in the elderly, providing timely care.

[0203] For examples consistent with the above embodiments, please refer to... Figure 6 , Figure 6 This is a functional unit block diagram of a sleep monitoring device that does not disturb sleep, as provided in an embodiment of this application. Figure 6 As shown, the undisturbed sleep monitoring device 60 includes: a first extraction unit 61, used to extract features from a first piezoelectric signal to obtain body movement features, wherein the first piezoelectric signal is obtained by superimposing multiple pre-processed original piezoelectric signals, and a single original piezoelectric signal includes multiple time points; an analysis unit 62, used to perform frequency domain analysis on a second piezoelectric signal, identify frequency components corresponding to peaks within the sleep heart rate correlation frequency band, extract frequency pairs corresponding to each time point to obtain a frequency set, wherein the frequency pairs include multiple preceding frequency components in a first frequency sequence and the frequency of each frequency component, wherein the first frequency sequence is obtained by arranging the frequency components in descending order according to amplitude intensity, and the second piezoelectric signal is the piezoelectric signal with the highest signal quality among the multiple original piezoelectric signals, wherein the signal quality is determined based on a preset number of sampling values ​​in a single original piezoelectric signal; a first determination unit 63, used to determine the heart rate features at each time point based on the frequency set; a second extraction unit 64, used to extract features from the acquired piezoresistive signal to obtain piezoresistive features; and a second determination unit 65, used to determine the sleep state based on the body movement features, the heart rate features at each time point, and the piezoresistive features.

[0204] In one possible embodiment, in extracting features from the first piezoelectric signal to obtain body motion features, the first extraction unit 61 is specifically configured to: extract features from the first piezoelectric signal to obtain body motion features, including: determining the peaks of the first piezoelectric signal to obtain a peak set; determining the peak mean of the peak set; determining a first number of abnormal peaks based on the peak set and the peak mean, wherein the abnormal peaks are used to indicate peaks in the peak set that are greater than a preset multiple of the peak mean; determining a body motion threshold based on the median of the first piezoelectric signal; and determining the body motion features based on the first number, the body motion threshold, and the first piezoelectric signal.

[0205] In one possible embodiment, in determining the body movement feature based on the first quantity, the body movement threshold, and the first piezoelectric signal, the first extraction unit 61 is further configured to: if the first quantity is detected to be non-zero, determine the target body movement value corresponding to the piezoelectric signal at each moment in the first piezoelectric signal based on the body movement threshold; determine the body movement feature based on the target body movement value; if the first quantity is detected to be zero, obtain an initial body movement value; and determine the body movement feature based on the initial body movement value.

[0206] In one possible embodiment, in determining the heart rate characteristics at each moment based on the set of frequencies, the first determining unit 63 is specifically configured to: determine the heart rate value at each moment based on each frequency pair in the set of frequencies; determine the signal quality of the piezoelectric signal at each moment in the second piezoelectric signal; and determine the heart rate characteristics based on the heart rate value and the signal quality.

[0207] In one possible embodiment, in determining the heart rate feature based on the heart rate value and the signal quality, the first determining unit 63 is further configured to: acquire a first heart rate value at the current moment and a second heart rate value at the previous moment; if it is detected that the signal quality of the piezoelectric signal at the current moment is greater than or equal to a preset signal quality, then determine a first weight and a second weight based on the signal quality of the piezoelectric signal at the current moment, wherein the first weight is the weight of the first heart rate value and the second weight is the weight of the second heart rate value; determine the heart rate feature at the current moment based on the first weight, the second weight, the first heart rate value, and the second heart rate value; if it is detected that the signal quality of the piezoelectric signal at the current moment is less than the preset signal quality, then acquire an initial heart rate value; and determine the initial heart rate value as the heart rate feature at the current moment.

[0208] In one possible embodiment, the piezoresistive feature includes a first feature and a second feature. In terms of feature extraction from the acquired piezoresistive signal to obtain the piezoresistive feature, the second extraction unit 64 is specifically used to: determine the first number of times the piezoresistive signal has a piezoresistive resistance less than a first preset threshold to obtain the first feature; determine the second number of times the piezoresistive signal has a piezoresistive resistance greater than a second preset threshold to obtain the second feature, wherein the second preset threshold is greater than the first preset threshold, and the second preset threshold and the first preset threshold are integer multiples of each other.

[0209] In one possible embodiment, in determining the sleep state based on the body movement characteristics, the heart rate characteristics at each moment, and the piezoresistive characteristics, the second determining unit 65 is further configured to: identify a sleep state based on the body movement characteristics to obtain a first state; identify a wakefulness state based on the body movement characteristics or the first characteristic to obtain a second state; identify sleep stages based on the body movement characteristics, the heart rate characteristics at each moment, the first characteristic, and the second characteristic to obtain a third state; and obtain the sleep state based on the first state, the second state, and the third state.

[0210] It is understood that since the method embodiments and the device embodiments are different presentations of the same technical concept, the content of the method embodiment section in this application should be adapted to the device embodiment section in a synchronous manner, and will not be repeated here.

[0211] In the case of using integrated units, please refer to Figure 7 , Figure 7 This is a functional unit block diagram of another sleep monitoring device provided in this application embodiment, such as... Figure 7 As shown, the undisturbed sleep monitoring device 60 includes a processing module 602 and a communication module 601. The processing module 602 controls and manages the operation of the undisturbed sleep monitoring device 60, for example, executing the steps of the first extraction unit 61, the analysis unit 62, the first determination unit 63, the second extraction unit 64, and the second determination unit 65, and / or performing other processes of the technology described herein. The communication module 601 is used for interaction between the undisturbed sleep monitoring device 60 and other devices.

[0212] Among them, such as Figure 7 As shown, the sleep undisturbed monitoring device 60 may also include a storage module 603, which is used to store the program code and data of the sleep undisturbed monitoring device 60.

[0213] The processing module 602 can be a processor or controller, such as a central processing unit (CPU), a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. The processor can also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0214] The communication module 601 can be a transceiver, RF circuit, or communication interface, etc. The storage module 603 can be a memory.

[0215] All relevant content in each scenario involved in the above method embodiments can be referenced from the functional descriptions of the corresponding functional modules, and will not be repeated here. The above-mentioned undisturbed sleep monitoring device 60 can perform the above-mentioned... Figure 2 The method for monitoring undisturbed sleep is shown.

[0216] Please see Figure 8 , Figure 8 This is a schematic diagram of the structure of an electronic device proposed in an embodiment of this application, as shown below. Figure 8 As shown, the electronic device 800 includes a processor 810, a memory 820, a communication interface 830, and one or more programs 821. The one or more programs 821 are stored in the memory and configured to be executed by the processor. When the program is executed, it includes some or all of the steps of any undisturbed sleep monitoring method described in the above method embodiments. The processor, memory, and communication interface are interconnected and complete communication between them.

[0217] The memory can be volatile memory such as Dynamic Random Access Memory (DRAM) or non-volatile memory such as a hard disk drive. The memory stores a set of executable program code, and the processor calls the executable program code stored in the memory to execute some or all of the steps of any of the sleep undisturbed monitoring methods described in the above embodiments of the sleep undisturbed monitoring method.

[0218] As can be seen, the electronic device 800 described in this application embodiment first extracts features from a first piezoelectric signal to obtain body motion features. The first piezoelectric signal is obtained by superimposing multiple pre-processed original piezoelectric signals, and a single original piezoelectric signal includes multiple moments. Then, frequency domain analysis is performed on the second piezoelectric signal to identify the frequency components corresponding to the peaks within the sleep heart rate correlation frequency band. Frequency pairs corresponding to each moment are extracted to obtain a frequency set. The frequency pairs include the preceding multiple frequency components in a first frequency sequence and the frequency of each frequency component. The first frequency sequence is obtained by arranging the frequency components in descending order according to amplitude intensity. The second piezoelectric signal is the piezoelectric signal with the highest signal quality among the multiple original piezoelectric signals. The signal quality is determined based on a preset number of sampling values ​​in a single original piezoelectric signal. Then, the heart rate features at each moment are determined based on the frequency set. Next, features are extracted from the acquired piezoresistive signal to obtain piezoresistive features. Finally, the sleep state is determined based on the body motion features, the heart rate features at each moment, and the piezoresistive features.

[0219] This application extracts body motion features from a first piezoelectric signal obtained by superimposing multiple original piezoelectric signals, which reduces single-channel errors and makes body motion judgment more stable. Furthermore, it selects a second piezoelectric signal with the best signal quality for frequency domain analysis, screening key frequency components to extract heart rate features, ensuring the accuracy of heart rate feature extraction. Through the above differentiated processing of multiple piezoelectric signals, the accuracy of feature extraction can be improved. Combined with the piezoresistive features extracted from the piezoresistive signal, a three-dimensional feature system of body motion, heart rate, and piezoresistive signals is formed, collaboratively judging sleep state, making sleep monitoring more comprehensive, and improving the reliability and accuracy of sleep state recognition.

[0220] This application also provides a computer storage medium storing a computer program for electronic data interchange, which causes a computer to perform some or all of the steps of any of the methods described in the above method embodiments, wherein the computer includes an electronic device.

[0221] This application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps of any of the methods described in the above method embodiments. The computer program product may be a software installation package, and the computer may include an electronic device.

[0222] It should be noted that, for the sake of simplicity, the aforementioned methods are described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are optional, and the actions and modules involved are not necessarily essential to this application.

[0223] In the above embodiments, the descriptions of each embodiment have their own emphasis. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0224] In the several embodiments provided in this application, it should be understood that the disclosed apparatus 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 system, 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 or other forms.

[0225] 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, depending on actual needs.

[0226] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software program module.

[0227] If the integrated unit is implemented as a software program module and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory 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 application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0228] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage device, which may include: a flash drive, a read-only memory, a random access memory, a magnetic disk, or an optical disk, etc.

[0229] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The above description of the embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for monitoring sleep without disturbing others, characterized in that, include: Feature extraction is performed on the first piezoelectric signal to obtain body motion features. The first piezoelectric signal is obtained by superimposing multiple preprocessed original piezoelectric signals. Each original piezoelectric signal includes multiple time points, and the multiple original piezoelectric signals are acquired by multiple different piezoelectric sensors within the same time window. The preprocessing includes differential processing and downsampling processing. Specifically, the peaks of the first piezoelectric signal are determined to obtain a peak set; the mean value of the peaks in the peak set is determined; and based on the peak set and the mean value, a first number of abnormal peaks is determined. The abnormal peaks are used to indicate that the peak set is greater than a preset multiple. The peak of the mean peak; a body movement threshold is determined based on the median of the first piezoelectric signal; the body movement feature is determined based on the first quantity, the body movement threshold, and the first piezoelectric signal; wherein, if the first quantity is detected to be non-zero, the target body movement value corresponding to the piezoelectric signal at each moment in the first piezoelectric signal is determined based on the body movement threshold; the body movement feature is determined based on the target body movement value; if the first quantity is detected to be zero, an initial body movement value is obtained, the initial body movement value being used to characterize no body movement event; the body movement feature is determined based on the initial body movement value, and the body movement feature is zero; Frequency domain analysis is performed on the second piezoelectric signal to identify the frequency components corresponding to the peaks within the sleep heart rate correlation frequency band. Frequency pairs corresponding to each time moment are extracted to obtain a frequency set. The frequency pairs include the preceding multiple frequency components in the first frequency sequence and the frequency of each frequency component. The first frequency sequence is obtained by arranging the frequency components in descending order according to their amplitude intensity. The second piezoelectric signal is the piezoelectric signal with the highest signal quality among the multiple original piezoelectric signals. The signal quality is determined based on a preset number of sampling values ​​in a single original piezoelectric signal. The heart rate characteristics at each moment are determined based on the frequency set; wherein, the heart rate value at each moment is determined by associating each frequency pair in the frequency set with integer multiples of the fundamental frequency and harmonics; the signal quality of the piezoelectric signal at each moment in the second piezoelectric signal is determined; and the heart rate characteristics are determined based on the heart rate value and the signal quality. Feature extraction is performed on the acquired piezoresistive signal to obtain piezoresistive features; Sleep state is determined based on the body movement characteristics, the heart rate characteristics at each moment, and the piezoresistive characteristics.

2. The method according to claim 1, characterized in that, Determining the heart rate characteristics based on the heart rate value and the signal quality includes: Get the first heart rate value at the current moment and the second heart rate value at the previous moment; If the signal quality of the piezoelectric signal at the current moment is detected to be greater than or equal to a preset signal quality, then a first weight and a second weight are determined based on the signal quality of the piezoelectric signal at the current moment. The first weight is the weight of the first heart rate value, and the second weight is the weight of the second heart rate value. The heart rate characteristics at the current moment are determined based on the first weight, the second weight, the first heart rate value, and the second heart rate value. If the signal quality of the piezoelectric signal at the current moment is detected to be less than the preset signal quality, then the initial heart rate value is obtained; The initial heart rate value is determined as the heart rate characteristic at the current moment.

3. The method according to any one of claims 1-2, characterized in that, The piezoresistive features include a first feature and a second feature. The step of extracting features from the acquired piezoresistive signal to obtain the piezoresistive features includes: The first feature is obtained by determining the first number of times that the piezoresistive signal has a piezoresistive resistance less than a first preset threshold. The second feature is obtained by determining the second number of times the piezoresistive signal is greater than the second preset threshold, wherein the second preset threshold is greater than the first preset threshold, and the second preset threshold and the first preset threshold are integer multiples of each other.

4. The method according to claim 3, characterized in that, Determining the sleep state based on the body movement characteristics, the heart rate characteristics at each moment, and the pressure resistance characteristics includes: Based on the described body movement characteristics, the sleep state is identified to obtain the first state; Based on the described body movement characteristics or the first characteristic, a waking state is identified to obtain a second state; Based on the body movement characteristics, the heart rate characteristics at each moment, the first characteristic, and the second characteristic, sleep stage identification is performed to obtain the third state; The sleep state is obtained based on the first state, the second state, and the third state.

5. A sleep monitoring device that does not disturb sleep, characterized in that, include: The first extraction unit is used to extract features from the first piezoelectric signal to obtain body motion features. The first piezoelectric signal is obtained by superimposing multiple preprocessed original piezoelectric signals. Each original piezoelectric signal includes multiple time points. The multiple original piezoelectric signals are acquired by multiple different piezoelectric sensors within the same time window. The preprocessing includes differential processing and downsampling processing. Specifically, the first extraction unit is further used to determine the peaks of the first piezoelectric signal to obtain a peak set; determine the peak mean of the peak set; and determine a first number of abnormal peaks based on the peak set and the peak mean. The abnormal peaks are used to indicate... The peak concentration is greater than a preset multiple of the peak mean; a body movement threshold is determined based on the median of the first piezoelectric signal; the body movement feature is determined based on the first quantity, the body movement threshold, and the first piezoelectric signal; wherein, if the first quantity is detected to be non-zero, the target body movement value corresponding to the piezoelectric signal at each moment in the first piezoelectric signal is determined based on the body movement threshold; the body movement feature is determined based on the target body movement value; if the first quantity is detected to be zero, an initial body movement value is obtained, the initial body movement value is used to characterize no body movement event; the body movement feature is determined based on the initial body movement value, and the body movement feature is zero; The analysis unit is used to perform frequency domain analysis on the second piezoelectric signal, identify the frequency components corresponding to the peaks within the sleep heart rate correlation frequency band, extract the frequency pairs corresponding to each time moment, and obtain a frequency set. The frequency pairs include the preceding multiple frequency components in the first frequency sequence and the frequency of each frequency component. The first frequency sequence is obtained by arranging the frequency components in descending order according to their amplitude intensity. The second piezoelectric signal is the piezoelectric signal with the highest signal quality among the multiple original piezoelectric signals. The signal quality is determined based on a preset number of sampling values ​​in a single original piezoelectric signal. The first determining unit is configured to determine the heart rate characteristics at each moment based on the frequency set; wherein, the first determining unit is further configured to determine the heart rate value at each moment based on each frequency pair in the frequency set by associating it with integer multiples of the fundamental frequency and harmonics; determine the signal quality of the piezoelectric signal at each moment in the second piezoelectric signal; and determine the heart rate characteristics based on the heart rate value and the signal quality. The second extraction unit is used to extract features from the acquired piezoresistive signal to obtain piezoresistive features; The second determining unit is used to determine the sleep state based on the body movement characteristics, the heart rate characteristics at each moment, and the piezoresistive characteristics.

6. An electronic device, characterized in that, The device includes: The device includes a memory, a processor, and executable program code stored in the memory and executable on the processor, wherein the processor executes the executable program code to perform the steps of the undisturbed sleep monitoring method as described in any one of claims 1-4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores executable program code, the executable program code including execution instructions for performing the steps of the undisturbed sleep monitoring method as described in any one of claims 1-4.

Images