A portable multi-modal non-visual biofeedback sleep aid device

Abstract

This invention relates to a sleep-aiding method and system based on heart rate variability biofeedback, belonging to the field of sleep health management technology. It solves the problem that existing sleep-aiding technologies, which rely on visual biofeedback and lack mechanisms for recognizing sleep trends and automatically exiting sleep patterns, are prone to overstimulation. The method includes: acquiring vital signs signals at multiple time steps and a sleep-aiding rhythm signal at the current time step; the vital signs signals include PPG sensor signals and IMU sensor signals; determining the signal quality index for each time step based on the PPG sensor signals from multiple time steps; calculating the body movement intensity at each time step based on the IMU sensor signals from multiple time steps; and adjusting the sleep-aiding rhythm signal based on the signal quality index and body movement intensity at the current time step or multiple time steps. This invention achieves full-process sleep management with light-free biofeedback, automatic sleep trend recognition, and nighttime awakening re-intervention, improving the effectiveness of sleep-aiding guidance.

Description

Technical Field

[0001] This invention relates to the field of sleep health management technology, and in particular to an assisted sleep method and system based on heart rate variability biofeedback. Background Technology

[0002] In the American Sleep Association's 2014 International Classification of Sleep Disorders, Third Edition, insomnia is defined as a persistent disturbance of sleep initiation, sleep duration, sleep continuity, or sleep quality at appropriate times and in appropriate environments, accompanied by impaired daytime functioning. The "2023 China Healthy Sleep White Paper" released by the Chinese Sleep Research Society shows that currently, 60.4% of people in China experience sleep-related symptoms.

[0003] However, healthcare professionals have limited intervention options for insomnia. Currently, first-line interventions mainly include short-term symptomatic medication and cognitive behavioral therapy (CBT-I) for insomnia. While web-based eCBT-I, developed with the internet, has reduced intervention costs, adherence remains unsatisfactory. More importantly, the efficacy of CBT for insomnia is limited. People with insomnia need more effective and low-risk interventions.

[0004] The pathological mechanisms of insomnia are related to abnormal emotions and arousal brain circuits. Abnormal autonomic nervous activity is also prevalent among insomnia patients. Based on the pathological mechanisms of insomnia symptoms, heart rate variability (HRV) biofeedback technology, which regulates the autonomic nervous system and modulates psychological stress, is an option for non-pharmacological intervention. However, current heart rate variability biofeedback technology for insomnia suffers from at least the following drawbacks: Using electrocardiogram (ECG) to collect heart rate data results in high costs and poor portability for wearable and deployment, making it difficult to apply in pre-sleep home settings; visual biofeedback is problematic because pre-sleep light stimulation can affect melatonin secretion and thus sleep, while auditory / tactile feedback with no or low light stimulation is more suitable, allowing participants to complete training in a supine position with their eyes closed and naturally transition to sleep; feedback based on a single HRV threshold is susceptible to motion artifacts, signal loss, and unstable breathing, leading to false feedback; moreover, the "awakening" of insomnia is not only reflected in HRV but also in multidimensional characteristics such as body movement, skin conductance (sympathetic excitation), and peripheral temperature, making it difficult for a single sensor to reliably characterize the pre-sleep state; the feedback method is limited and cannot comprehensively cover the entire process of "guided breathing—reward fading—sleep withdrawal—nighttime awakening re-intervention"; and the lack of individualized parameters makes it difficult to adapt fixed breathing rhythms to all insomniacs.

[0005] Therefore, there is an urgent need for a sleep-aiding method and system based on heart rate variability biofeedback. This solves the problem that existing sleep-aiding technologies, which use visual biofeedback and lack recognition of sleep trends and automatic exit mechanisms, are prone to overstimulation. This method achieves full-process sleep management with no light biofeedback, automatic recognition of sleep trends, and re-intervention upon nighttime awakenings, thereby improving the effectiveness of sleep-aiding guidance. Summary of the Invention

[0006] In view of the above analysis, the present invention aims to provide a sleep aid method and system based on heart rate variability biofeedback, in order to solve the problem that existing sleep aid technologies, which use visual biofeedback and lack recognition of sleep onset trends and automatic exit mechanisms, are prone to overstimulation.

[0007] On one hand, embodiments of the present invention provide a sleep-aiding method based on heart rate variability biofeedback, comprising: Acquire vital signs signals at multiple time steps and sleep-inducing rhythm signals at the current time step; wherein, the vital signs signals include PPG sensor signals and IMU sensor signals; the multiple time steps include the current time step; Based on the PPG sensor signals at multiple time steps, determine the signal quality index corresponding to each time step; Based on the IMU sensor signals at multiple time steps, calculate the body motion intensity corresponding to each time step; The sleep-aid rhythm signal is adjusted based on the signal quality index and the body movement intensity at the current time step or multiple time steps.

[0008] Further, based on the PPG sensor signals at multiple time steps, a signal quality index corresponding to each time step is determined, including: Peak point identification is performed on the PPG sensor signals at multiple time steps to obtain a number of corresponding PPG signal peak points; Based on the PPG sensor signals at multiple time steps and several PPG signal peaks, determine the morphological stability score and peak stability score corresponding to each time step; Bandpass filtering is performed on the PPG sensor signals at multiple time steps to obtain PPG filtered signals corresponding to multiple time steps; Based on the PPG sensor signals and the PPG filtered signals at multiple time steps, calculate the bandpass energy fraction corresponding to each time step; The corresponding signal quality index is calculated based on the morphological stability score, the peak stability score, and the bandpass energy score for each time step.

[0009] Further, based on the PPG sensor signals at multiple time steps and all PPG signal peaks, the morphological stability score and peak stability score corresponding to each time step are determined, including: For each PPG signal peak, the corresponding local waveform segment and peak value are extracted from the PPG sensor signals at multiple time steps; For each time step, a corresponding quality index time window is determined, and for each local waveform segment corresponding to the quality index time window, the corresponding Pearson correlation coefficient is calculated. For each time step, the morphological stability score corresponding to the time step is calculated based on all the Pearson correlation coefficients corresponding to the corresponding quality index time window; For each time step, the peak stability score corresponding to the time step is calculated based on all the peak values ​​corresponding to the corresponding quality index time window.

[0010] Further, based on the IMU sensor signals at multiple time steps, the body motion intensity corresponding to each time step is calculated, including: Based on the IMU sensor signal at each time step, calculate the degravity dynamic acceleration corresponding to each time step; The body motion intensity corresponding to each time step is calculated based on the degravation dynamic acceleration at multiple time steps.

[0011] Further, the sleep-aid rhythm signal is adjusted based on the signal quality index and the body movement intensity at the current time step, including: The signal quality index at the current time step is validated to obtain a first validity result, and the body motion intensity at the current time step is validated to obtain a second validity result. If the first validity result or the second validity result is passed, then the heart rate consistency index of the current time step is calculated based on the PPG sensor signal of the current time step. The sleep-aid rhythm signal is adjusted based on the heart rate consistency index at the current time step.

[0012] Further, the signal quality index at the current time step is validated to obtain a first validity result, and the body motion intensity at the current time step is validated to obtain a second validity result, including: If the signal quality index is not less than the quality index threshold, then the first validity result is determined to be passed; If the intensity of the body movement is not greater than the body movement threshold, then the second validity result is determined to be passed.

[0013] Furthermore, adjusting the sleep-aiding sound signal based on the signal quality index and the body movement intensity also includes: If both the first validity result and the second validity result are unsuccessful, the heart rate consistency index is frozen, and the freezing duration is recorded. If the freezing duration reaches the freezing threshold, the sleep-aiding rhythm signal is maintained.

[0014] Further, the sleep-aid rhythm signal is adjusted based on the signal quality index and the body movement intensity at multiple time steps, including: PPG coverage and body motion coverage are calculated based on the signal quality index and body motion intensity at multiple time steps. If both the PPG coverage rate and the body movement coverage rate are not less than the corresponding coverage threshold, then the physical sign baseline corresponding to the current time step is determined based on the physical sign signals of the multiple time steps. Based on the vital signs signals corresponding to multiple time steps and the vital signs baseline of the current time step, determine whether the current time step is in a state of nighttime wakefulness; If the person is in the aforementioned nighttime awakening state, then the sleep-aiding sound signal will be activated.

[0015] Furthermore, the vital signs baseline includes heart rate baseline, heart rate variability baseline, and body motion baseline; Based on the vital signs signals and the vital signs baseline corresponding to multiple time steps, it is determined whether the current time step is in a state of nighttime wakefulness, including: The heart rate state corresponding to the current time step is determined based on the PPG sensor signals corresponding to multiple time steps and the heart rate baseline corresponding to the current time step. The variability state corresponding to the current time step is determined based on the PPG sensor signals corresponding to multiple time steps and the heart rate variability baseline corresponding to the current time step. The body motion state corresponding to the current time step is determined based on the IMU sensor signals corresponding to the multiple time steps and the body motion baseline corresponding to the current time step; Based on the heart rate status, the variability status, and the body movement status, determine whether the current time step is in the night-wake state.

[0016] On the other hand, embodiments of the present invention provide a sleep aid system based on heart rate variability biofeedback, comprising: The signal acquisition module is used to acquire vital sign signals at multiple time steps and sleep aid rhythm signals at the current time step; wherein, the vital sign signals include PPG sensor signals and IMU sensor signals; the multiple time steps include the current time step; The vital signs analysis module is used to determine the signal quality index corresponding to each time step based on the PPG sensor signals of multiple time steps. The vital signs analysis module is also used to calculate the body motion intensity corresponding to each time step based on the IMU sensor signals of multiple time steps; The rhythm adjustment module is used to adjust the sleep-aid rhythm signal based on the signal quality index and the body movement intensity of the current time step or multiple time steps.

[0017] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects: This system utilizes a PPG (photoplethysmography) sensor to acquire biofeedback, reducing the impact of light exposure and making it suitable for various sleep conditions. By incorporating signal quality index and body movement intensity, it assesses PPG signal quality and discriminates IMU body movements, reducing motion artifacts and signal loss on feedback accuracy and improving the robustness of feedback decisions. Adjustments to the sleep-aiding rhythm signal are made based on the signal quality index and body movement intensity at the current time step, maintaining or reducing the signal to guide breathing and fade out the rhythm. Continuous monitoring of sleep patterns is achieved through multiple time steps of signal quality index and body movement intensity, enabling timely detection of night awakenings and adjustments to the sleep-aiding rhythm signal. This allows for re-intervention and guidance to fall back asleep after a night awakening, achieving phased, intensity-based, and adaptive intervention throughout the entire sleep process, thus enhancing the effectiveness of sleep guidance.

[0018] In this invention, the above-described technical solutions can be combined with each other to achieve more preferred combinations. Other features and advantages of this invention will be set forth in the following description, and some advantages may become apparent from the description or be learned by practicing the invention. The objects and other advantages of this invention can be realized and obtained from what is particularly pointed out in the description and drawings. Attached Figure Description

[0019] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. Figure 1 This is a flowchart illustrating a sleep-aiding method based on heart rate variability biofeedback in an embodiment of the present invention. Figure 2 This is a schematic diagram of the main modules of a sleep aid system based on heart rate variability biofeedback in an embodiment of the present invention. Detailed Implementation

[0020] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.

[0021] Step S1: Acquire vital signs signals from multiple time steps and sleep-inducing rhythm signals from the current time step; wherein, the vital signs signals include PPG sensor signals and IMU sensor signals; the multiple time steps include the current time step.

[0022] The system acquires vital signs signals at multiple time steps and sleep-aiding sound signals at the current time step. In this embodiment, the vital signs signals include PPG sensor signals and IMU sensor signals, and the multiple time steps include the current time step. By analyzing the changes in PPG sensor signals and IMU sensor signals, the sleep-aiding sound signals are adjusted in real time to guide and assist sleep.

[0023] In this embodiment, the pulse interval (IBI, approximating the RR interval) of the person requiring sleep assistance (i.e., the user) is collected using a PPG sensor. The PPG sensor signal, which is the signal collected by the PPG sensor, is used to characterize the user's IBI interval (i.e., the time interval between two heartbeats, typically a few tenths of a millisecond). The PPG sensor refers to a sensor using photoplethysmography (PPG) technology; preferably, multi-wavelength PPG (green / red / infrared light) is used to improve signal quality and adaptability. Body movement information is collected using an IMU sensor. The IMU sensor signal, which is the signal collected by the IMU sensor, is used to characterize the intensity of the user's body movements; preferably, a three-axis accelerometer and / or gyroscope is used. In this embodiment, Bluetooth communication is used to transmit the PPG sensor and IMU sensor signals.

[0024] Step S2: Determine the signal quality index corresponding to each time step based on the PPG sensor signals of the multiple time steps.

[0025] Based on the PPG sensor signals at multiple time steps, the corresponding signal quality index is calculated for each time step, which is used to eliminate false feedback and as the basis for adjusting the sleep aid rhythm signal. In this embodiment, step S2 includes steps S21-S25.

[0026] Step S21: Perform peak point identification on the PPG sensor signals of multiple time steps to obtain a number of corresponding PPG signal peak points.

[0027] Peak point identification is performed on PPG sensor signals at multiple time steps to obtain several corresponding PPG signal peak points. In this embodiment, the step size of each time step is 1 second, and a peak detection algorithm (such as the findpeaks function in MATLAB) is used to identify peak points in PPG sensor signals at multiple consecutive time steps to obtain several PPG signal peak points. In other embodiments, peak point identification can also be performed by combining main wave peak detection, feature point marking, signal preprocessing and noise filtering, and multimodal feature fusion; no limitation is made here.

[0028] Step S22: Based on the PPG sensor signals of multiple time steps and several PPG signal peaks, determine the morphological stability score and peak stability score corresponding to each time step.

[0029] Based on PPG sensor signals at multiple time steps and several PPG signal peaks, the morphological stability score and peak stability score corresponding to each time step are determined. In this embodiment, a sliding window method is used to determine the morphological stability score and peak stability score corresponding to the last time step of each quality index time window using PPG sensor signals at multiple time steps and several PPG signal peaks. For example, the time window length of the quality index time window is 10s.

[0030] Step S22 specifically includes steps S221-S224.

[0031] Step S221: For each PPG signal peak, extract the corresponding local waveform segment and peak value from the PPG sensor signals of multiple time steps.

[0032] For each PPG signal peak, the corresponding local waveform segment and peak value are extracted from the PPG sensor signals at multiple time steps. In this embodiment, local waveform segment extraction is performed based on the peak neighbor step size. For each PPG signal peak, the band within the peak neighbor step size is extracted from the PPG sensor signals at multiple time steps as the local waveform segment corresponding to that PPG signal peak. For example, if the time corresponding to a certain PPG signal peak is 12:00:00:0200 milliseconds, and the peak neighbor step size includes the previous neighbor step size (e.g., 0.15 seconds, i.e., 150 milliseconds) and the next neighbor step size (e.g., 0.25 seconds, i.e., 250 milliseconds), then the band between 12:00:00:00:0050 milliseconds and 12:00:00:00:0450 milliseconds is extracted as the local waveform segment corresponding to that PPG signal peak. Furthermore, to improve data accuracy, the local waveform segments of all PPG signal peaks are normalized. In this embodiment, DC is removed and the amplitude is normalized according to the PPG signal peaks so that amplitude differences do not dominate the correlation (i.e., Pearson correlation coefficient).

[0033] In this embodiment, peak value extraction is performed based on peak height, that is, the peak height of each PPG signal peak is taken as the corresponding peak value. In other embodiments, the amplitude of each PPG signal peak can also be taken as the corresponding peak value, or all amplitudes can be normalized and then used as the corresponding peak value.

[0034] Furthermore, to improve the effectiveness of the PPG sensor signal, outlier detection is performed based on the number of PPG signal peaks corresponding to each quality index time window. If the number of PPG signal peaks corresponding to each quality index time window is less than the minimum number of peaks (e.g., 5), the PPG sensor signal for that quality index time window is considered insufficient for evaluating the heart rate consistency index, the signal quality index is forced to 0, and the freeze strategy in step S44 is triggered.

[0035] Step S222: For each time step, determine the corresponding quality index time window, and for each local waveform segment corresponding to the quality index time window, calculate the corresponding Pearson correlation coefficient.

[0036] For each time step, a corresponding quality index time window is determined. For each local waveform segment corresponding to the quality index time window, the corresponding Pearson correlation coefficient is calculated. In this embodiment, for each time step, this time step is taken as the last time step of the quality index time window, and the corresponding quality index time window is determined. Based on the template waveform segment, the Pearson correlation coefficient corresponding to each local waveform segment of the quality index time window is calculated. The template waveform segment is determined based on all local waveform segments of the quality index time window. For example, the local waveform segment with the middle peak value among all local waveform segments is taken as the template waveform segment. , This represents the k-th local waveform segment. This indicates a local waveform segment with the peak value centered.

[0037] For example, the step size of each time step is 1 second, and the time window length of the quality index time window is 10 seconds. For the 20th time step, the time range of the corresponding quality index time window is from the 11th time step to the 20th time step. If the quality index time window includes 5 local waveform segments, and the peak values ​​of each local waveform segment are 0.1, 0.2, 0.3, 0.4, and 0.5, then the local waveform segment corresponding to the median peak value of 0.3 is taken as the template waveform segment. The Pearson correlation coefficients of the five local waveform segments and the template waveform segment are calculated sequentially to characterize the waveform similarity between each local waveform segment and the template waveform segment. For example, the k-th local waveform segment... ( (where n is the amplitude of the i-th sampling point and n is the number of sampling points), template waveform segment ( (where n is the amplitude of the i-th sampling point and n is the number of sampling points), and the Pearson correlation coefficient corresponding to the k-th local waveform segment. In the formula, The average amplitude of the k-th local waveform segment. This represents the average amplitude of the template waveform segment.

[0038] Step S223: For each time step, calculate the morphological stability score corresponding to the time step based on all the Pearson correlation coefficients corresponding to the corresponding quality index time window.

[0039] For each time step, the morphological stability score is calculated based on all Pearson correlation coefficients corresponding to the corresponding quality index time window. In this embodiment, the mean of all Pearson correlation coefficients for the corresponding quality index time window is used as the morphological stability score for that time step.

[0040] Furthermore, to eliminate the influence of dimensions, all Pearson correlation coefficients are mapped to... Then, the mean of all Pearson correlation coefficients after all mappings is used as the morphological stability score for that time step. ,in, This represents the mapped Pearson correlation coefficient corresponding to the k-th local waveform segment. This indicates the number of local waveform segments within the quality index time window.

[0041] Step S224: For each time step, calculate the peak stability score corresponding to the time step based on all the peak values ​​corresponding to the corresponding quality index time window.

[0042] For each time step, a peak stability score is calculated based on all peak values ​​within the corresponding quality index time window. In this embodiment, for each time step, a peak stability score is calculated based on the standard deviation and mean of the peak values ​​of all PPG signal peaks within the corresponding quality index time window. For example, the peak stability score... ,in, This represents the peak value of the k-th PPG signal peak. This represents the standard deviation of the peak values ​​of all PPG signal peaks within the quality index time window. This represents the mean of the peak values ​​of all PPG signal peaks within the quality index time window. To represent non-zero terms, avoid division by zero, for example... .

[0043] Furthermore, to eliminate the influence of dimensions, the peak stability score is mapped, and subsequent calculations or judgments are performed using the mapped peak stability score. For example, the mapped peak stability score... In the formula, This represents the upper limit of the mapped peak stability score, making the peak stability score more sensitive to peak value changes. If the peak value changes at different PPG signal peaks increase, the mapped peak stability score will rapidly drop to near 0. For example... The mapped peak stability score is at Interval.

[0044] Step S23: Perform bandpass filtering on the PPG sensor signals of the multiple time steps to obtain the PPG filtered signals corresponding to the multiple time steps.

[0045] Bandpass filtering is performed on PPG sensor signals at multiple time steps to obtain PPG filtered signals corresponding to multiple time steps. In this embodiment, the PPG sensor signals are filtered by setting an upper and lower limit of the bandpass frequency to evaluate whether the pulse signal dominates the PPG signal. For example, based on the common heart rate range of 42-240 beats / minute in sleep scenarios, an upper limit of the bandpass frequency is set. Lower limit of bandpass frequency Bandpass filtering is performed on the PPG sensor signals at multiple time steps to obtain the corresponding PPG filtered signals at multiple time steps. The timing information of the PPG sensor signal and the PPG filter signal is consistent.

[0046] Step S24: Calculate the bandpass energy fraction corresponding to each time step based on the PPG sensor signals and the PPG filtered signals at multiple time steps.

[0047] Based on PPG sensor signals and PPG filtered signals from multiple time steps, the bandpass energy fraction for each time step is calculated. In this embodiment, for each time step, the total fluctuation energy is calculated based on the PPG sensor signals from all time steps within the corresponding quality index time window, and the bandpass energy is calculated based on the PPG filtered signals from all time steps. The ratio of the total fluctuation energy to the bandpass energy is used as the bandpass energy fraction for that time step. For example, the total fluctuation energy... Energy Bandpass energy fraction ,in, Let be the energy of the PPG sensor signal at the t-th sampling point. Let be the energy of the PPG filtered signal at the t-th sampling point. This indicates the number of sampling points corresponding to the time window of the quality index. To represent non-zero terms, avoid division by zero, for example... .

[0048] Step S25: Calculate the corresponding signal quality index based on the morphological stability score, peak stability score, and bandpass energy score for each time step.

[0049] The signal quality index is calculated based on the morphological stability score, peak stability score, and bandpass energy score for each time step. In this embodiment, for each time step, the weighted sum of the morphological stability score, peak stability score, and bandpass energy score is used as the signal quality index for that time step. For example, the signal quality index for the m-th time step... ,in, , , Let represent the morphological stability score, the mapped peak stability score, and the bandpass energy score at the m-th time step, respectively. , , Represents weights, for example , , .

[0050] Furthermore, to reduce instantaneous fluctuations, a first-order low-pass smoothing is performed on the signal quality index at each time step. The smoothed signal quality index is then used for subsequent calculations or assessments. For example, the smoothed signal quality index at the m-th time step... ,in, Let be the smoothed signal quality index at the (m-1)th time step. For example, smoothing coefficients .

[0051] Step S3: Calculate the body motion intensity corresponding to each time step based on the IMU sensor signals of multiple time steps.

[0052] Based on the IMU sensor signals at multiple time steps, the body motion intensity corresponding to each time step is calculated, specifically including steps S31-S32.

[0053] Step S31: Calculate the gravity dynamic acceleration corresponding to each time step based on the IMU sensor signal at each time step.

[0054] Based on the IMU sensor signals at each time step, the degravity dynamic acceleration corresponding to each time step is calculated. In this embodiment, according to the formula... , Calculate the degravation dynamic acceleration corresponding to each time step, where, This indicates the acceleration magnitude corresponding to that time step. ( , , (Acceleration along the x-axis, y-axis, and z-axis, in that order). This represents a low-frequency gravity trend estimate, within the corresponding gravity trend time window at that time step (e.g., Among the acceleration magnitudes of all time steps, the median magnitude is selected as the low-frequency gravity trend estimate for that time step. Alternatively, a low-pass method can be used to determine the low-frequency gravity trend estimate for that time step. Represents gravity. .

[0055] Step S32: Calculate the body motion intensity corresponding to each time step based on the degravity dynamic acceleration of the multiple time steps.

[0056] The body motion intensity for each time step is calculated based on the degravation dynamic acceleration at multiple time steps. In this embodiment, for each time step, the root mean square of the degravation dynamic acceleration for all time steps included in the corresponding mass index time window is calculated as the body motion intensity for that time step. .

[0057] For example, the m-th time step includes Each sampling point corresponds to a gravity trend time window. include Each sampling point corresponds to a quality index time window including... Sampling points (usually) Greater than ), the acceleration modulus of the t-th sampling point ( , , The x-axis, y-axis, and z-axis accelerations at the t-th sampling point are obtained sequentially (this can be obtained from the motion sensor signal). This time step... The average acceleration magnitude of each sampling point is used as the acceleration magnitude at that time step. Take the gravity trend time window corresponding to that time step. of The median of the acceleration modulus at each sampling point is used as the low-frequency gravity trend estimate for that time step. The dynamic acceleration at that time step Degravity dynamic acceleration Physical activity intensity (N is the number of time steps included in the quality index time window). In other embodiments, this time step is taken. The acceleration magnitude of each sampling point is used as the multiple acceleration magnitudes at that time step. Low-frequency gravity trend estimation at this time step Dynamic acceleration Degravity dynamic acceleration Physical activity intensity .

[0058] Step S4: Adjust the sleep-aid rhythm signal according to the signal quality index and body movement intensity of the current time step or multiple time steps.

[0059] The sleep-aiding sound signal is adjusted based on the signal quality index and body movement intensity at the current time step or multiple time steps. In this embodiment, the sleep-aiding method includes adjusting the sleep-aiding sound signal based on the signal quality index and body movement intensity at the current time step, or adjusting the sleep-aiding sound signal based on the signal quality index and body movement intensity at multiple time steps. For example, the adjustment method is determined based on the signal state of the sleep-aiding sound signal. For instance, if the sleep-aiding sound signal is in an active state, it is adjusted based on the signal quality index and body movement intensity at the current time step; if the sleep-aiding sound signal is in a deactivated state, it is adjusted based on the signal quality index and body movement intensity at multiple time steps.

[0060] Based on the signal quality index and body movement intensity at the current time step, the sleep-aiding rhythm signal is adjusted, including steps S41-S43.

[0061] Step S41: Verify the validity of the signal quality index at the current time step to obtain a first validity result, and verify the validity of the body motion intensity at the current time step to obtain a second validity result.

[0062] The signal quality index at the current time step is validated to obtain a first validity result, and the body motion intensity at the current time step is validated to obtain a second validity result, including steps A and B.

[0063] Step A: If the signal quality index is not less than the quality index threshold, then the first validity result is determined to be passed.

[0064] If the signal quality index is not less than the quality index threshold, the first validity result is determined to be passed. In this embodiment, the average signal quality index of all time steps collected by the user on the current day or multiple days is used as the quality index threshold. In other embodiments, the quality index threshold can also be set empirically, for example... .

[0065] Step B: If the intensity of the body movement is not greater than the body movement threshold, then the second validity result is determined to be passed.

[0066] If the body motion intensity is not greater than the body motion threshold, the second validity result is determined to be passed. In this embodiment, the average value of the gravity dynamic acceleration collected by the user at all time steps on the same day or multiple days is used as the body motion threshold. In other embodiments, the body movement threshold can be set empirically. ,For example .

[0067] Step S42: If the first validity result or the second validity result is passed, then calculate the heart rate consistency index of the current time step based on the PPG sensor signal of the current time step.

[0068] If either the first or second validity result is passed, the PPG sensor signal is considered reliable, and the heart rate consistency index is calculated based on the PPG sensor signal at the current time step. First, a Fast Fourier Transform (FFT) is performed on the PPG sensor signal at the current time step to obtain the corresponding power spectral density. Second, a filtering frequency range is determined based on the general laws of human respiration; for example, the filtering frequency range is 0.04 Hz to 0.26 Hz. All power values ​​within the filtering frequency range are then counted as the total power. Next, the main peak frequency is determined from the PPG sensor signal at the current time step. Centered on the main peak frequency, all power within the adjacent frequency band is statistically analyzed and taken as the main peak power. The neighborhood frequency band can be set empirically, for example, 0.015, which is... All power within the range is taken as the main peak power. Finally, based on the total power... and peak power Calculate the heart rate consistency index .

[0069] Furthermore, to improve data reliability and stability, after artifact removal, data interpolation, and DC component removal of the PPG sensor signal, a Hann-Window is applied to prevent spectral leakage. The PPG sensor signal after applying the Hann-Window is then used for a Fast Fourier Transform. In this embodiment, artifact removal is based on local average values; for example, the interval mean of the IBI interval over eight consecutive time steps is calculated. Compare each IBI interval within these 8 time steps with the mean interval; if it exceeds... If the IBI interval is abnormal, it is considered an artifact and removed. In other embodiments, artifact removal can also be based on local average values, upper and lower limit intervals (e.g., upper limit interval 1200 ms, lower limit interval 300 ms), or continuous jumps (e.g., the relative error between two adjacent IBI intervals is greater than 20%, or the absolute difference is greater than 250 ms). For the PPG sensor signal after artifact removal, cubic spline interpolation is used to fill in gaps in the data. The interpolated PPG sensor signal is then processed to remove the DC component to eliminate constant components in the signal and ensure the accuracy of subsequent frequency domain analysis.

[0070] By verifying signal quality and the effectiveness of body movement signals, PPG sensor signals affected by body movement are eliminated, and abnormal IBI values ​​are screened out to avoid false feedback and improve the stability and effectiveness of sleep-aiding sound signals.

[0071] Step S43: Adjust the sleep-aid rhythm signal according to the heart rate consistency index of the current time step.

[0072] Based on the heart rate consistency index at the current time step, the sleep-aid rhythm signal is adjusted. Specifically, if the heart rate consistency index at the current time step is not less than the consistency threshold, the sleep-aid rhythm signal is positively adjusted; if the heart rate consistency index at the current time step is less than the consistency threshold, the sleep-aid rhythm signal is negatively adjusted. In this embodiment, the consistency threshold... Based on experience, for example ,like This indicates that the main peak power ratio is high within the filtered frequency range, and the heart rate shows obvious consistency with breathing. When the user is in a consistent or relaxed state, positive feedback regulation is applied to the sleep aid sound signal. For example, the sleep aid sound signal is reduced so that the sleep aid sound signal in the next time step is lower than that in the current time step. The weakened sound of the sleep aid sound signal gives the user a sense of security by creating a sense of distance, and the sleep aid sound signal based on the absence of light guides the user to fall asleep.

[0073] Furthermore, a consistency time threshold can be set. Adjusting the sleep aid rhythm signal based on the heart rate consistency index at the current time step also includes: if the heart rate consistency index at the current time step is not less than the consistency threshold, then the duration for which the heart rate consistency index is greater than the consistency threshold is counted. If the consistency time threshold is reached, then the sleep aid rhythm signal is reduced. For example, if the consistency time threshold is 10 time steps, and the heart rate consistency index at the current time step is not less than the consistency threshold, and the heart rate consistency index at the previous 9 time steps is not less than the consistency threshold, then the sleep aid rhythm signal is reduced.

[0074] In this embodiment, positive feedback adjustment of the sleep-aid sound signal includes gradually decreasing the volume of the sleep-aid sound according to a smoothing function (e.g., decreasing the total volume by 1% per second), or decreasing it exponentially, or switching the music type of the sleep-aid sound (e.g., switching the volume of the ocean wave sound to a low volume b1). Negative feedback adjustment of the sleep-aid sound signal includes keeping the volume of the sleep-aid sound constant, or gradually increasing it according to a smoothing function (e.g., increasing the existing volume by 1% per second), or switching the music type of the sleep-aid sound (e.g., switching the ocean wave volume to a high volume b2). For example, the sleep-aid sound signal uses two audio signals: a breathing cue signal and a progress enhancement signal. For instance, the sound of ocean waves is used as the breathing cue signal. Based on the comparison between the heart rate consistency index and the consistency threshold at each time step, the volume of the ocean wave sound is adjusted. If the heart rate consistency index is not greater than the consistency threshold, it is considered that the user is not relaxed, and the volume of the ocean wave sound is increased to prompt the user to breathe at the frequency of the ocean wave sound, guiding the user to gradually enter a relaxed state where the heart rate consistency index is greater than the consistency threshold. Background music is used as the progress enhancement signal. If the heart rate consistency index is greater than the consistency threshold, it is considered that the user is relaxed and gradually falling asleep, and the volume of the ocean wave sound and background music is decreased.

[0075] Furthermore, the duration of sleep assistance can be set. After the sleep state score calculated in step S5 reaches the sleep state limit, the volume of the sleep aid sound signal is reduced at equal intervals according to the duration of sleep assistance until it is turned off. For example, if the duration of sleep assistance is 16 minutes, then when the sleep state score reaches the sleep state limit, the volume of the sleep aid sound based on the current time step is reduced at equal intervals until it is turned off.

[0076] In other embodiments, adjusting the sleep-aid rhythm signal based on the signal quality index and body movement intensity at the current time step further includes adjusting the sleep-aid rhythm signal based on the heart rate consistency index and body movement intensity at the current time step. Specifically, if the heart rate consistency index at the current time step is not less than the consistency threshold and the body movement intensity is less than the body movement threshold, the sleep-aid rhythm signal is reduced; if the heart rate consistency index at the current time step is less than the consistency threshold or the body movement intensity is not less than the body movement threshold, the sleep-aid rhythm signal is enhanced.

[0077] Furthermore, after step S41, the following steps are also included: Step S44: If both the first validity result and the second validity result are unsuccessful, then freeze the heart rate consistency index and count the freeze duration; if the freeze duration reaches the freeze threshold, then maintain the sleep aid rhythm signal.

[0078] If both the first and second validity results are unsuccessful (i.e., the signal quality index is less than the quality index threshold and the body motion intensity is greater than the body motion threshold), indicating low reliability of the PPG sensor signal, then the heart rate consistency index is frozen. This means the heart rate consistency index from the previous time step is used as the heart rate consistency index for the current time step. Simultaneously, the freeze duration is calculated as the total duration of time steps where both the first and second validity results are unsuccessful. If the freeze duration reaches a freeze threshold (e.g., a freeze threshold...), then... If the sleep aid sound signal is maintained, the sleep aid sound signal in the next time step will be the same as the sleep aid sound signal in the current time step, thus avoiding false feedback.

[0079] Furthermore, if the freeze duration does not reach the freeze threshold, then referring to step S43, the sleep aid rhythm signal is adjusted according to the heart rate consistency index of the current time step (i.e., the heart rate consistency index of the previous time step obtained by freezing the heart rate consistency index).

[0080] The sleep-aid rhythm signal is adjusted based on the signal quality index and the body movement intensity at multiple time steps, including steps S45-S48.

[0081] Step S45: Calculate PPG coverage and motion coverage based on the signal quality index and motion intensity of the multiple time steps.

[0082] Based on the signal quality index and body motion intensity at multiple time steps, PPG coverage and body motion coverage are calculated. In this embodiment, a sliding window approach is used to calculate PPG coverage and body motion coverage based on the signal quality index and body motion intensity at multiple time steps corresponding to each trend assessment time window. These are then used as the PPG coverage and body motion coverage corresponding to the last time step of the trend assessment time window. For example, the trend assessment time window ( The time window length is 10 minutes.

[0083] The total duration of time steps in which the signal quality index of multiple time steps corresponding to the statistical trend assessment time window is not less than the quality index threshold is taken as the PPG coverage duration. The percentage of PPG coverage time within the trend assessment time window is called the PPG coverage rate. The total duration of time steps within which the body movement intensity does not exceed the body movement threshold, corresponding to multiple time steps of the statistical trend assessment time window, is taken as the body movement coverage duration. The percentage of physical activity coverage time within the trend assessment time window is the physical activity coverage rate. .

[0084] Step S46: If both the PPG coverage rate and the body movement coverage rate are not less than the corresponding coverage threshold, then determine the vital sign baseline corresponding to the current time step based on the vital sign signals of the multiple time steps.

[0085] If both PPG coverage and body motion coverage are not less than their corresponding coverage thresholds, then the vital sign baseline for the current time step is determined based on vital sign signals from multiple time steps. This vital sign baseline includes the heart rate baseline, heart rate variability baseline, and body motion baseline. For example, the coverage threshold corresponding to PPG coverage... The coverage threshold corresponding to the physical coverage rate ,like and Then, the baseline of vital signs corresponding to the current time step is determined based on the vital signs signals from multiple time steps.

[0086] In this embodiment, the vital signs baseline includes the heart rate baseline, heart rate variability baseline, and body movement baseline. The mean heart rate, mean heart rate variability, and mean body movement intensity across multiple time steps corresponding to a single duration window of the sleep aid sound signal are used as the heart rate baseline, heart rate variability baseline, and body movement baseline for the current time step. The single duration window of the sleep aid sound signal refers to the time range from when the user actively activates the sleep aid sound signal to the current time step, or the time range from when the sleep aid sound signal is activated based on step S48 to the current time step. That is, within the initial sleep guidance cycle or within the sleep guidance cycle after intervention during nighttime awakenings, the vital signs baseline is determined independently for each cycle.

[0087] For example, the time difference between two adjacent PPG signal peaks is taken as the IBI interval, and the mean of all IBI intervals for multiple time steps corresponding to a single duration window of the sleep-inducing rhythm signal is taken as the IBI interval. Calculate the mean heart rate, i.e., the baseline heart rate. The root mean square of the difference between adjacent IBI intervals was used as the indicator of heart rate variability. The root mean square of the differences between all adjacent IBI intervals across multiple time steps corresponding to a single duration window of the sleep-inducing rhythm signal was used as the mean of heart rate variability, i.e., the baseline of heart rate variability. ,in This refers to the number of IBI intervals corresponding to the trend assessment time window. , Let represent the IBI intervals corresponding to the t-th and t-1-th PPG signal peaks, respectively; the average of body movement intensity across multiple time steps corresponding to a single duration window of the sleep-inducing rhythm signal is used as the body movement baseline. In other embodiments, the body movement threshold can also be used as the body movement baseline.

[0088] Step S47: Based on the vital signs signals corresponding to the multiple time steps and the vital signs baseline corresponding to the current time step, determine whether the current time step is in a state of night awakening; Based on the vital signs and baseline values ​​corresponding to multiple time steps, it is determined whether the current time step is a nighttime awakening state. In this embodiment, a sliding window approach is used, with each nighttime awakening check time window (e.g., The corresponding vital signs and the baseline of vital signs at the current time step are used to determine whether the current time step is in a state of wakefulness at night, specifically including steps S471-S474.

[0089] Step S471: Determine the heart rate state corresponding to the current time step based on the PPG sensor signals corresponding to the multiple time steps and the heart rate baseline corresponding to the current time step.

[0090] First, for the current time step, the heart rate is determined based on the PPG sensor signals from multiple time steps within the corresponding nighttime wake-up check window. Within the corresponding nighttime wake-up check window, multiple heart rates are extracted, and the average of all heart rates is used as the heart rate for the current time step. In other embodiments, the median of all heart rates can be taken as the heart rate at the current time step.

[0091] Secondly, compare the current heart rate with the baseline heart rate to determine the corresponding heart rate status. For example, if... If, then the heart rate state at the current time step is in an upward heart rate state; if Then the heart rate state at the current time step is a steady heart rate state; where, For example, the heart rate variability coefficient. .

[0092] Step S472: Determine the variability state corresponding to the current time step based on the PPG sensor signals corresponding to the multiple time steps and the heart rate variability baseline corresponding to the current time step.

[0093] First, for the current time step, the heart rate variability is determined based on the PPG sensor signals corresponding to multiple time steps within the corresponding nighttime wake-up check time window. The root mean square of the differences between all adjacent IBI intervals within the corresponding multiple time steps within the corresponding nighttime wake-up check time window is used as the heart rate variability for the current time step. .

[0094] Secondly, compare the heart rate variability at the current time step with the baseline heart rate variability to determine the corresponding variability state. For example, if... If, then the variability state at the current time step is a decreasing variability state; if If the variability state at the current time step is increasing, then the variability state is increasing; where, For example, the coefficient of variation fluctuation. .

[0095] Step S473: Determine the body motion state corresponding to the current time step based on the IMU sensor signals corresponding to the multiple time steps and the body motion baseline corresponding to the current time step.

[0096] First, for the current time step, based on the IMU sensor signals corresponding to multiple time steps within the corresponding night wake-up check time window, the body motion intensity for each time step is determined with reference to step S3. .

[0097] Secondly, the body motion intensity at each time step is compared with the body motion baseline at the current time step. The total duration of time steps where the body motion intensity is no greater than the body motion baseline is counted and used as the body motion stability duration. ,in Calculate the percentage of body motion stability time within the nighttime wake-up check time window. If the proportion is not less than the body motion coefficient ( If the proportion is less than the body movement coefficient, then the current time step's body movement state is an increased body movement state; If the body motion state at the current time step is stable, then the body motion state at the current time step is a stable body motion state; where, For example, the coefficient of variation fluctuation. .

[0098] Step S474: Based on the heart rate status, the variability status, and the body movement status, determine whether the current time step is in the night-wake state.

[0099] Based on heart rate status, variability status, and body movement status, it is determined whether the current time step is in a nighttime awakening state. In this embodiment, if the heart rate status of the current time step is in an increasing heart rate state, the variability status is in a decreasing variability state, and the body movement status is in a state of increased body movement, then it is determined that the current time step is in a nighttime awakening state.

[0100] In other embodiments, the vital signs signals also include EDA signals and skin temperature signals, and the vital signs baselines also include EDA baselines and skin temperature baselines. Based on the vital signs signals corresponding to multiple time steps and the vital signs baselines corresponding to the current time step, it is determined whether the current time step is in a nocturnal awakening state. It also includes using the average values ​​of EDA signals and skin temperatures of multiple time steps corresponding to a single duration window of the sleep aid sound signal as EDA baselines and skin temperature baselines, respectively, to determine whether the EDA signal of the current time step exceeds the EDA baseline and whether the skin temperature signal exceeds the skin temperature baseline. When the heart rate state of the current time step is in an increasing heart rate state, the variability state is in a decreasing variability state, and the body movement state is in an increased body movement state, and the EDA signal exceeds the EDA baseline and the skin temperature signal exceeds the skin temperature baseline, it is determined that the current time step is in a nocturnal awakening state.

[0101] Step S48: If the person is in the nighttime awakening state, then activate the sleep aid sound signal.

[0102] If the current time step is a nighttime awakening state, a nighttime awakening re-intervention is triggered, initiating a sleep-aiding audio signal to guide the user back to sleep. In other embodiments, the sleep-aiding audio signal also includes voice reminders, using low-frequency, short phrases to prompt the user to fall back asleep.

[0103] To avoid frequent activation of sleep aid signals due to momentary nighttime awakenings, which could disrupt the user's sleep, this embodiment also sets a nighttime awakening duration threshold and an intervention duration threshold. If a user is awake at night, the sleep aid signal is activated. The embodiment also includes determining whether the continuous duration of the nighttime awakening reaches the nighttime awakening duration threshold; if so, the sleep aid signal is activated. Furthermore, the embodiment determines the duration of the sleep aid signal; if the duration reaches the intervention duration threshold, the sleep aid signal is deactivated. For example, the nighttime awakening duration threshold... Once the continuous duration from the first time step in the nighttime awakening state to the current time step in the nighttime awakening state reaches the nighttime awakening duration threshold, the sleep aid sound signal is activated, and the duration of the sleep aid sound signal is recorded. Once the intervention duration threshold is reached, the sleep aid sound signal is forcibly turned off.

[0104] Furthermore, to reduce the computational load, a first update cycle can be set. For each first update cycle, referring to steps S45-S48, the heart rate, heart rate variability, and body movement intensity of the first update cycle are calculated, and the corresponding baseline vital signs are compared to determine whether the person is in a nocturnal awakening state.

[0105] In this embodiment, after step S45, the method further includes: Step S49: If the PPG coverage rate or the body movement coverage rate is less than the corresponding coverage threshold, then maintain the sleep-aiding sound signal.

[0106] If the PPG coverage or body movement coverage is less than the corresponding coverage threshold, the sleep-inducing sound signal will be maintained, and intervention will not be triggered if nighttime awakenings are not detected.

[0107] Furthermore, the sleep-aiding method based on heart rate variability biofeedback provided in this embodiment also includes: S5: Determine the corresponding sleep onset trend based on the vital signs signals at multiple time steps.

[0108] A sliding window approach is used to calculate the sleep state score for each trend assessment time window. This score serves as the sleep state score for the last time step within that trend assessment window. The sleep state scores for all time steps within a given day constitute the sleep trend for that day. Based on the sleep trend, body movement thresholds, quality index thresholds, and coverage thresholds can be adjusted. This approach can also be used to assist in assessing sleep effectiveness. For example, the average body movement intensity for all time steps within the sleep trend whose sleep state scores fall within the threshold range can be used as a new body movement threshold for the next sleep induction session.

[0109] In this embodiment, a second update cycle is set, and a sleep state score is calculated once for each second update cycle, for example... The mean heart rate, mean heart rate variability, mean body movement intensity, and mean heart rate consistency of all time steps within the second update cycle are used as the corresponding periodic heart rate, periodic variability, periodic body movement, and periodic consistency for that second update cycle.

[0110] Furthermore, the sleep state score includes heart rate score, variability score, body movement score, and consistency score. The sleep state score is obtained by weighted fusion of these scores. In the formula , , , In order, they are heart rate score, variability score, body movement score, and consistency score. , , , The corresponding weights are listed in order, for example... , , , .

[0111] In this embodiment, heart rate scoring is achieved using a formula. Calculate, where, This represents the slope of the HR linear regression. Represents a mapping, if ,but ,like ,but ,like ,but .in, The calculation is performed through the following steps: Select time steps within the trend assessment time window where the signal quality index is not less than the quality index threshold; within the selected time steps, extract the corresponding multiple heart rates. A linear regression was performed on all extracted heart rates to obtain the corresponding linear regression slope. ,in The unit is bpm / min, and a negative value indicates a decrease in heart rate.

[0112] Variability score heart rate score via formula Calculate, where, This represents the slope of the RMSSD linear regression. Represents a mapping, if ,but ,like ,but ,like ,but .in, The calculation is performed through the following steps: extracting heart rate variability for each time step within the trend assessment time window, and then performing linear regression fitting on all extracted heart rate variability to obtain the corresponding linear regression slope. ,in The unit is ms / min, and a positive value indicates an increase.

[0113] Body movement score includes body movement rate and postural fluctuation, where body movement rate is the most important factor. attitude fluctuation Body movement score is calculated based on body movement rate and posture fluctuation. In the formula, This indicates the total duration of time steps within which the body movement intensity does not exceed the body movement baseline within the trend assessment time window. , These represent the standard deviations of the roll angle and pitch angle for all sampling points within the trend assessment time window, respectively. Pitch angle , , , The accelerations are, in order, the x-axis, y-axis, and z-axis.

[0114] Consistency scoring is achieved through a formula. Calculate, where, This indicates the percentage of time during which the heart rate consistency index is not less than the consistency threshold within the trend assessment time window. ( This indicates the cumulative duration during which the heart rate consistency index is not less than the consistency threshold.

[0115] Furthermore, if the PPG coverage rate or physical activity coverage rate is less than the corresponding coverage threshold, the sleep state score of the previous second update cycle will be used as the sleep state score of the current second update cycle to ensure the validity of the sleep state score.

[0116] Through the above steps, PPG sensor signals and IMU sensor signals are monitored in real time to analyze the user's sleep state. By using audio with sleep-inducing effects or optional tactile rhythms (i.e. sleep-inducing sound signals), the user can optimize autonomic nervous system regulation and heart rate variability through biofeedback, thereby improving sleep.

[0117] This invention provides an assisted sleep system based on heart rate variability biofeedback, such as... Figure 2 As shown, it includes: The signal acquisition module is used to acquire vital sign signals at multiple time steps and sleep aid rhythm signals at the current time step; wherein, the vital sign signals include PPG sensor signals and IMU sensor signals; the multiple time steps include the current time step; The vital signs analysis module is used to determine the signal quality index corresponding to each time step based on the PPG sensor signals of multiple time steps. The vital signs analysis module is also used to calculate the body motion intensity corresponding to each time step based on the IMU sensor signals of multiple time steps; The rhythm adjustment module is used to adjust the sleep-aid rhythm signal based on the signal quality index and the body movement intensity of the current time step or multiple time steps.

[0118] Furthermore, the vital signs analysis module is also used to determine the corresponding sleep onset trend based on vital signs signals from multiple time steps.

[0119] This system is programmed into a program or software that can run on a mobile device. Taking a mobile app as an example, the system's operation is illustrated as follows: Before going to sleep, the user opens the mobile app and wears a wristband (including a PPG sensor and an IMU sensor). The user breathes according to the rhythm guided by the sleep-aiding sound signal (e.g., 5 seconds of inhalation followed by 5 seconds of exhalation, or 4 seconds of inhalation followed by 6 seconds of exhalation, which is more conducive to improving anxiety). Under this breathing rhythm, the heart rate variability exhibits characteristic periodic changes. The wristband collects PPG and IMU sensor signals and transmits them to the mobile app via Bluetooth. Based on the system disclosed in this embodiment, the mobile app adjusts the sleep-aiding sound signal in real time using closed-loop feedback logic. When the user reaches the set consistency threshold, it is considered that the user is in a relaxed sleep state, and the sleep-aiding sound signal is gradually reduced and faded out (e.g., the volume of the sleep-aiding sound signal is reduced at equal intervals over 16 minutes). The PPG and IMU sensor signals are continuously collected. If the user is found to be awake at night, the sleep-aiding sound signal is activated again to guide the user's breathing to fall back asleep.

[0120] The above-described method and system embodiments are based on the same principles, and their related aspects can be referenced from each other to achieve the same technical effects. For specific implementation processes, please refer to the foregoing embodiments, which will not be repeated here.

[0121] In summary, the sleep-aiding method and system based on heart rate variability biofeedback according to embodiments of the present invention has at least one of the following beneficial effects: This method utilizes PPG sensors to acquire biofeedback, avoiding the influence of light. By introducing signal quality index and body movement intensity, it achieves PPG signal quality assessment and IMU body movement discrimination, reducing the impact of motion artifacts and signal loss on feedback accuracy and improving the robustness of feedback decisions. By adjusting the sleep aid rhythm signal based on the signal quality index and body movement intensity at the current time step, it maintains or reduces the current sleep aid rhythm signal, completes guided breathing, and fades out the sleep aid rhythm signal. Through signal quality index and body movement intensity at multiple time steps, it continuously monitors sleep status, promptly detects night awakenings, adjusts the sleep aid rhythm signal, and completes night awakening intervention to guide the patient back to sleep. This achieves phased, intensity-based, and adaptive full-process sleep intervention, improving the effectiveness of assisted sleep guidance.

[0122] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0123] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A sleep-aiding method based on heart rate variability biofeedback, characterized in that, include: Acquire vital signs signals at multiple time steps and sleep-inducing rhythm signals at the current time step; wherein, the vital signs signals include PPG sensor signals and IMU sensor signals; the multiple time steps include the current time step; Based on the PPG sensor signals at multiple time steps, determine the signal quality index corresponding to each time step; Based on the IMU sensor signals at multiple time steps, calculate the body motion intensity corresponding to each time step; The sleep-aid rhythm signal is adjusted based on the signal quality index and the body movement intensity at the current time step or multiple time steps.

2. The method according to claim 1, characterized in that, Based on the PPG sensor signals at multiple time steps, determine the signal quality index corresponding to each time step, including: Peak point identification is performed on the PPG sensor signals at multiple time steps to obtain a number of corresponding PPG signal peak points; Based on the PPG sensor signals at multiple time steps and several PPG signal peaks, determine the morphological stability score and peak stability score corresponding to each time step; Bandpass filtering is performed on the PPG sensor signals at multiple time steps to obtain PPG filtered signals corresponding to multiple time steps; Based on the PPG sensor signals and the PPG filtered signals at multiple time steps, calculate the bandpass energy fraction corresponding to each time step; The corresponding signal quality index is calculated based on the morphological stability score, the peak stability score, and the bandpass energy score for each time step.

3. The method according to claim 2, characterized in that, Based on the PPG sensor signals at multiple time steps and all PPG signal peaks, determine the morphological stability score and peak stability score corresponding to each time step, including: For each PPG signal peak, the corresponding local waveform segment and peak value are extracted from the PPG sensor signals at multiple time steps; For each time step, a corresponding quality index time window is determined, and for each local waveform segment corresponding to the quality index time window, the corresponding Pearson correlation coefficient is calculated. For each time step, the morphological stability score corresponding to the time step is calculated based on all the Pearson correlation coefficients corresponding to the corresponding quality index time window; For each time step, the peak stability score corresponding to the time step is calculated based on all the peak values ​​corresponding to the corresponding quality index time window.

4. The method according to claim 1, characterized in that, Based on the IMU sensor signals at multiple time steps, the body motion intensity corresponding to each time step is calculated, including: Based on the IMU sensor signal at each time step, calculate the degravity dynamic acceleration corresponding to each time step; The body motion intensity corresponding to each time step is calculated based on the degravation dynamic acceleration at multiple time steps.

5. The method according to claim 1, characterized in that, The sleep-aid rhythm signal is adjusted based on the signal quality index and body movement intensity at the current time step, including: The signal quality index at the current time step is validated to obtain a first validity result, and the body motion intensity at the current time step is validated to obtain a second validity result. If the first validity result or the second validity result is passed, then the heart rate consistency index of the current time step is calculated based on the PPG sensor signal of the current time step. The sleep-aid rhythm signal is adjusted based on the heart rate consistency index at the current time step.

6. The method according to claim 5, characterized in that, The signal quality index at the current time step is validated to obtain a first validity result, and the body motion intensity at the current time step is validated to obtain a second validity result, including: If the signal quality index is not less than the quality index threshold, then the first validity result is determined to be passed; If the intensity of the body movement is not greater than the body movement threshold, then the second validity result is determined to be passed.

7. The method according to claim 5, characterized in that, Adjusting the sleep-aid rhythm signal based on the signal quality index and the body movement intensity further includes: If both the first validity result and the second validity result are unsuccessful, the heart rate consistency index is frozen, and the freezing duration is recorded. If the freezing duration reaches the freezing threshold, the sleep-aiding rhythm signal is maintained.

8. The method according to claim 1, characterized in that, The sleep-aid rhythm signal is adjusted based on the signal quality index and body movement intensity at multiple time steps, including: PPG coverage and body motion coverage are calculated based on the signal quality index and body motion intensity at multiple time steps. If both the PPG coverage rate and the body movement coverage rate are not less than the corresponding coverage threshold, then the physical sign baseline corresponding to the current time step is determined based on the physical sign signals of the multiple time steps. Based on the vital signs signals corresponding to multiple time steps and the vital signs baseline of the current time step, determine whether the current time step is in a state of nighttime wakefulness; If the person is in the aforementioned nighttime awakening state, then the sleep-aiding sound signal will be activated.

9. The method according to claim 8, characterized in that, The baseline vital signs include the heart rate baseline, heart rate variability baseline, and body movement baseline; Based on the vital signs signals and the vital signs baseline corresponding to multiple time steps, it is determined whether the current time step is in a state of nighttime wakefulness, including: The heart rate state corresponding to the current time step is determined based on the PPG sensor signals corresponding to multiple time steps and the heart rate baseline corresponding to the current time step. The variability state corresponding to the current time step is determined based on the PPG sensor signals corresponding to multiple time steps and the heart rate variability baseline corresponding to the current time step. The body motion state corresponding to the current time step is determined based on the IMU sensor signals corresponding to the multiple time steps and the body motion baseline corresponding to the current time step; Based on the heart rate status, the variability status, and the body movement status, determine whether the current time step is in the night-wake state.

10. A sleep aid system based on heart rate variability biofeedback, characterized in that, include: The signal acquisition module is used to acquire vital sign signals at multiple time steps and sleep aid rhythm signals at the current time step; wherein, the vital sign signals include PPG sensor signals and IMU sensor signals; the multiple time steps include the current time step; The vital signs analysis module is used to determine the signal quality index corresponding to each time step based on the PPG sensor signals of multiple time steps. The vital signs analysis module is also used to calculate the body motion intensity corresponding to each time step based on the IMU sensor signals of multiple time steps; The rhythm adjustment module is used to adjust the sleep-aid rhythm signal based on the signal quality index and the body movement intensity of the current time step or multiple time steps.

Images