Adaptive adjustment method and device based on sleep-aiding parameters, equipment and medium

Abstract

This invention discloses an adaptive adjustment method, device, equipment, and medium based on sleep-aid parameters, relating to the field of artificial intelligence technology. The method includes: collecting multimodal physiological characteristic data of a user and obtaining sleep evaluation data within a preset period; inputting the multimodal physiological characteristic data into a preset sleep state recognition model to obtain sleep stage states and abnormal fluctuation states; generating a joint sleep-aid parameter configuration including an initial threshold based on the sleep stage states and a preset parameter mapping rule, to control the operation of the sleep-aid device; when an abnormal fluctuation state is detected, triggering a parameter conflict suppression mechanism to safely adjust the joint sleep-aid parameter configuration; and finally calculating a sleep improvement score based on the sleep evaluation data using a preset periodic evaluation model, and updating the initial threshold of the joint sleep-aid parameter configuration based on the score. This effectively improves the targeting, safety, and sleep improvement effect of sleep-aid intervention.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to an adaptive adjustment method, device, equipment and medium based on sleep aid parameters. Background Technology

[0002] With the fast pace of modern life, sleep disorders have become a widespread health problem affecting the public. Quality sleep is crucial for physiological repair and mental well-being. Currently, the gold standard for diagnosing sleep disorders in clinical practice is polysomnography (PSG), which accurately determines sleep stages by collecting various physiological signals such as electroencephalograms (EEG), electrooculograms (EOG), and electromyograms (EMG). However, PSG examinations typically require patients to undergo testing in medical institutions, which is not only costly and cumbersome, but also involves complex equipment, making continuous and routine monitoring and intervention difficult in a home environment.

[0003] To improve sleep quality, existing sleep aids typically employ relatively fixed stimulation protocols, such as pre-set durations of sleep-aiding audio playback, constant light color temperature, or magnetic stimulation at preset frequencies. This static approach to sleep aids has significant limitations: First, the sleep process is highly dynamic, encompassing different stages such as initiation of sleep, light sleep, and deep sleep, and individual physiological states vary, making it difficult for fixed stimulation parameters to adapt to the user's real-time changing physiological needs. Second, it lacks a real-time feedback loop. When users experience abnormal fluctuations such as frequent body movement, abnormally elevated heart rate (HR), or decreased blood oxygen saturation (SpO2), existing static sleep aid systems often fail to recognize and respond safely in a timely manner, potentially causing the sleep-aiding stimulation to interfere with the user or even trigger physiological discomfort.

[0004] Furthermore, existing sleep aid solutions mostly focus on immediate intervention during a single sleep session, lacking a quantitative evaluation mechanism for long-term sleep-aiding effects. Most devices cannot integrate subjective sleep feedback with objective physiological indicators, making it difficult to adjust and optimize sleep aid strategies through periodic data accumulation. This results in sleep aid parameters being in a state of "blind adjustment" for extended periods. For users suffering from chronic insomnia, this fails to create a targeted, personalized intervention plan.

[0005] In summary, existing sleep aid technologies still have technological gaps in real-time perception of multimodal physiological signals, accurate identification of sleep states, adaptive adjustment of sleep aid parameters, and updating of intervention strategies based on long-term evaluation. This results in room for improvement in the efficiency and safety of sleep aid interventions. Therefore, developing an intelligent technology solution capable of achieving closed-loop adaptive adjustment of sleep aid parameters based on real-time physiological feedback has become an urgent issue to be addressed in this field. Summary of the Invention

[0006] This invention provides an adaptive adjustment method, device, equipment, and medium based on sleep-aid parameters, which solves the problem that existing sleep-aid intervention programs lack a closed-loop adaptive adjustment mechanism based on real-time physiological feedback, resulting in low efficiency and insufficient safety of individualized sleep aids.

[0007] The first aspect of this invention proposes an adaptive adjustment method based on sleep-aid parameters, comprising: Collect users' multimodal physiological characteristic data and obtain users' sleep evaluation data within a preset period; Multimodal physiological feature data is input into a preset sleep state recognition model to obtain sleep stage states and abnormal fluctuation states; Based on the sleep stage state, a combined sleep aid parameter configuration containing an initial threshold is generated through preset parameter mapping rules. The combined sleep aid parameter configuration is used to control the operation of the sleep aid device. When an abnormal fluctuation is detected, the parameter conflict suppression mechanism is triggered to safely adjust the combined sleep aid parameter configuration and control the sleep aid device to work according to the adjusted combined sleep aid parameter configuration. Based on sleep evaluation data, a sleep improvement score is calculated using a preset periodic evaluation model, and the initial thresholds of the combined sleep aid parameters are updated based on the sleep improvement score.

[0008] In one alternative implementation, the multimodal physiological data includes at least body movement data, heart rate data, respiratory data, electroencephalogram (EEG) data, and blood oxygenation data. Multimodal physiological feature data is input into a pre-defined sleep state recognition model to obtain sleep stage states and abnormal fluctuation states, including: Signal denoising and feature alignment are performed on body motion data, heart rate data, respiratory data, and blood oxygen data, and multi-dimensional temporal variation features are extracted. The temporal variation features are input into a sleep state recognition model pre-deployed on the local edge and built on a lightweight deep learning network. The sleep stage state is calculated in real time through the temporal convolutional layer and classification fully connected layer inside the model. The system monitors body movement frequency in body movement data, the magnitude of heart rate increase in heart rate data, and the magnitude of blood oxygen decrease in blood oxygen data in parallel. When the body movement frequency, the magnitude of heart rate increase, or the magnitude of blood oxygen decrease exceeds the safety baseline set for the individual, it is determined to be in an abnormal fluctuation state.

[0009] In one optional implementation, the sleep stages include at least the initial sleep state, light sleep state, and deep sleep state; the combined sleep-aid parameter configuration includes at least audio stimulation parameters, light stimulation parameters, and magnetic stimulation parameters; Based on sleep stage status, and through preset parameter mapping rules, a combined sleep-aid parameter configuration including an initial threshold is generated, including: When the sleep stage is in the state of falling asleep, audio stimulation parameters including low-frequency sleep-inducing frequency, fixed slow-changing rhythm and decreasing volume are generated, as well as light stimulation parameters including gradually decreasing low brightness and sleep-inducing warm color temperature. When the sleep stage is light sleep or deep sleep, magnetic stimulation parameters including magnetic stimulation intensity, magnetic stimulation frequency and single duration are generated within the preset brain nerve safety threshold range. The playback volume of the audio stimulation parameters is gradually reduced from light sleep to deep sleep until it stops playing completely.

[0010] In one alternative implementation, when an abnormal fluctuation is detected, a parameter conflict suppression mechanism is triggered to safely adjust the combined sleep aid parameter configuration, including: Analyze the specific anomaly types of abnormal fluctuation states; When the specific abnormality type is an abnormal drop in blood oxygen, it is determined that there is a physiological safety conflict in the configuration of the combined sleep aid parameters. The output of audio stimulation parameters, light stimulation parameters and magnetic stimulation parameters is immediately and forcibly suspended, and health reminder parameters for triggering external alarms are generated. When the specific abnormality is a state of frequent body movement or an abnormally high heart rate, the output intensity of the magnetic stimulation parameters is dynamically reduced according to a preset attenuation ratio, and the audio stimulation parameters are switched to soothing white noise with calming and calming effects to inhibit the deterioration of the state and help the patient fall back asleep.

[0011] In one alternative implementation, the sleep evaluation data includes subjective feedback data and sleep index statistics extracted from sleep monitoring records. The sleep index statistics include sleep latency duration, total number of nighttime awakenings, percentage of deep sleep, and sleep continuity index. Based on sleep assessment data, a sleep improvement score is calculated using a pre-set periodic assessment model, including: Sleep index statistics and subjective feedback data are input into a periodic evaluation model constructed based on a multi-dimensional weighted evaluation algorithm to calculate objective physiological improvement scores and subjective psychological satisfaction scores, respectively. The sleep improvement score, which characterizes the overall improvement in sleep quality, is obtained by weighting and fusing the objective physiological improvement score and the subjective psychological satisfaction score according to the pre-assigned weight coefficients.

[0012] In one alternative implementation, the initial thresholds of the combined sleep aid parameter configuration are updated based on the sleep improvement score, including: Determine whether the sleep improvement score is lower than the preset expected improvement threshold; If the sleep improvement score is lower than the expected improvement threshold, then the weak points of the sleep indicator statistics will be diagnosed, and the target sleep indicator with the lowest score will be identified. Based on the target sleep indicators and combined with the individual sleep-wake patterns of users, the system uses a preset circadian rhythm mathematical model to recalculate and dynamically update the initial thresholds of the intervention time interval, stimulation intensity change rate, and trigger time baseline value in the combined sleep aid parameter configuration, in order to generate a revised intervention plan.

[0013] In one optional implementation, based on the sleep stage state and through preset parameter mapping rules, a joint sleep-aid parameter configuration including an initial threshold is generated, which further includes: Sleep slow-wave signals are extracted in real time from continuously collected EEG data, and the real-time phase sequence of the sleep slow-wave signals is calculated using a phase-locked loop algorithm. Using real-time phase timing as a closed-loop feedback trigger condition, when the sleep slow wave signal is detected to reach a specific upward phase window, the stimulation pulse output corresponding to the joint sleep aid parameter configuration is precisely and synchronously triggered to achieve positive enhancement modulation of the sleep slow wave amplitude.

[0014] A second aspect of this invention provides an adaptive adjustment device based on sleep-aid parameters, comprising: The data acquisition module is used to collect the user's multimodal physiological characteristic data and obtain the user's sleep evaluation data within a preset period; The state recognition module is used to input multimodal physiological feature data into a preset sleep state recognition model to obtain sleep stage states and abnormal fluctuation states. The parameter generation module is used to generate a combined sleep aid parameter configuration, including an initial threshold, based on the sleep stage state and through preset parameter mapping rules. The combined sleep aid parameter configuration is used to control the operation of the sleep aid device. The safety suppression module is used to trigger the parameter conflict suppression mechanism to safely adjust the combined sleep aid parameter configuration when abnormal fluctuations are detected, and to control the sleep aid device to work according to the adjusted combined sleep aid parameter configuration. The evaluation update module is used to calculate a sleep improvement score based on sleep evaluation data using a preset periodic evaluation model, and to update the initial threshold of the combined sleep aid parameters based on the sleep improvement score.

[0015] A third aspect of the present invention provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by at least one processor. When the computer program is executed by at least one processor, it implements the adaptive adjustment method based on sleep-aid parameters as proposed in the foregoing embodiments.

[0016] The fourth aspect of this invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the adaptive adjustment method based on sleep-aid parameters as proposed in the foregoing embodiments.

[0017] Beneficial effects: The embodiments of the present invention provide an adaptive adjustment method, device, equipment and medium based on sleep aid parameters. By collecting users' multimodal physiological characteristic data and combining it with sleep evaluation data, a closed-loop control logic is constructed, which includes sleep state identification, dynamic generation of joint parameters, safe suppression of abnormal fluctuations and periodic evaluation updates, thereby improving the pertinence of sleep aid intervention. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the electronic device structure of the hardware operating environment involved in the embodiments of the present invention; Figure 2 This is a flowchart illustrating the steps of an adaptive adjustment method based on sleep-aid parameters provided in an embodiment of the present invention. Figure 3 This is a schematic diagram of the functional modules of an adaptive adjustment device based on sleep-aid parameters provided in an embodiment of the present invention. Detailed Implementation

[0019] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0020] Reference Figure 1 This is a schematic diagram of the electronic device structure of the hardware operating environment involved in the embodiments of the present invention.

[0021] like Figure 1As shown, the electronic device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless-fidelity (Wi-Fi) interface. The memory 1005 may be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM), such as a disk drive. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0022] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0023] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a data storage module, a network communication module, a user interface module, and computer programs.

[0024] exist Figure 1 In the electronic device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the electronic device of the present invention can be set in the electronic device, and the electronic device calls the adaptive adjustment device based on sleep-aid parameters stored in the memory 1005 through the processor 1001, and executes the adaptive adjustment method based on sleep-aid parameters provided in the embodiment of the present invention.

[0025] Combined with appendix Figure 2 This invention provides an adaptive adjustment method based on sleep-aid parameters, comprising the following steps: S101 collects the user's multimodal physiological characteristic data and obtains the user's sleep evaluation data within a preset period.

[0026] In one alternative implementation, the multimodal physiological data includes at least body movement data, heart rate data, respiratory data, electroencephalogram (EEG) data, and blood oxygen saturation data. To achieve multidimensional perception of the multimodal physiological data, multiple sensor nodes deployed on the user's body surface or around the bed are wirelessly or wiredly connected via a communication interface.

[0027] S1011. Acquire multimodal physiological characteristic data. Specifically, acquire the data stream generated in real time by the multimodal sensor: S10111. Collect body motion data. A three-axis accelerometer built into the smart mattress or wearable device acquires the user's acceleration changes in three-dimensional space. This body motion data not only characterizes the user's physical movements during sleep, such as turning over and limb twitching, but can also be further analyzed using algorithms to determine the user's sleep posture (e.g., supine, side-lying) and the frequency of these changes. The sampling rate is typically set between 50 Hz and 100 Hz to capture brief bursts of body motion, thereby accurately quantifying the user's nocturnal restlessness.

[0028] S10112. Acquire heart rate and blood oxygen data. This is done using a photoplethysmography (PPG) sensor. Specifically, the PPG sensor emits a beam of light of a specific wavelength (such as red and infrared light) to illuminate the user's skin blood vessels. By detecting changes in the intensity of reflected or transmitted light, the heart rate data is obtained by calculating the cardiac cycle. Simultaneously, based on the differences in the absorption rates of oxyhemoglobin and deoxyhemoglobin in the blood to different wavelengths of light, real-time blood oxygen saturation data is calculated.

[0029] S10113. Collect respiratory data. This can be done using a piezoelectric breathing chest cuff or by extracting the respiratory baseline drift from the PPG signals mentioned above. By analyzing the time interval between peaks and troughs, this data can accurately characterize respiratory rate, respiratory depth, and potential apnea events, providing data support for subsequent cross-validation of abnormal blood oxygenation.

[0030] S10114. Acquire Electroencephalogram (EEG) Data. Electroencephalogram (EEG) signals are acquired using dry or wet electrodes attached to the user's forehead or mastoid process. EEG data contains rich frequency band information and is a core basis for determining sleep stages. For example, Delta waves (0.5 Hz to 4 Hz) are extracted to assess deep sleep quality, and Alpha waves (8 Hz to 13 Hz) are extracted to determine the user's wakefulness and relaxation state.

[0031] S1012. Obtain the user's sleep evaluation data within a preset period.

[0032] In one alternative implementation, the sleep evaluation data includes subjective feedback data and sleep index statistics extracted from sleep monitoring records. The sleep index statistics include sleep latency duration, total number of awakenings during the night, percentage of deep sleep, and sleep continuity index.

[0033] S10121. Extract sleep index statistics. After the end of each natural day's sleep, perform offline statistical analysis on the continuous multimodal physiological characteristic data of that night. Calculate the time span from when the user lies down to prepare for sleep until the EEG data first shows a continuous light sleep waveform as the sleep latency duration; by tracking continuous sleep cycles, count the number of times the user transitions from sleep to wakefulness during the entire sleep period as the total number of nighttime awakenings; calculate the ratio of the total duration of deep sleep to the total sleep duration as the proportion of deep sleep; and calculate the variance of sleep segment distribution based on a Markov chain transition probability model to obtain the sleep continuity index. For example, if there are frequent high-frequency reverse transitions between deep sleep and light sleep, the model will output a lower continuity index. Smooth the above data according to a preset period (e.g., 7 days or 14 days) to form sleep index statistics with reference value.

[0034] S10122. Obtain subjective feedback data. A standardized sleep quality questionnaire is pushed to users after they wake up via a human-computer interaction interface. The questionnaire can use a Visual Analogue Scale to collect users' self-ratings of the ease of falling asleep (e.g., perception of the number of times they toss and turn), overall sleep depth, and morning mental recovery status (e.g., residual fatigue), forming subjective feedback data.

[0035] S102, input the multimodal physiological feature data into the preset sleep state recognition model to obtain the sleep stage state and abnormal fluctuation state.

[0036] In one optional implementation, multimodal physiological feature data is input into a preset sleep state recognition model to obtain sleep stage states and abnormal fluctuation states, including the following steps: S1021. Perform signal denoising and feature alignment on body motion data, heart rate data, respiratory data, and blood oxygen data, and extract multi-dimensional temporal variation features.

[0037] S10211, Signal Denoising. For EEG data, a bandpass filter is used to retain the frequency band signal from 0.5 Hz to 35 Hz, eliminating high-frequency electromyography interference and power frequency noise; for heart rate and respiration data, a moving average filtering algorithm is applied to smooth the curves, eliminate slow baseline drift, and ensure the stability of the signal baseline; for body motion data, a high-pass filter is applied to filter out the constant acceleration component caused by static gravity, thus retaining only the dynamic acceleration changes that characterize actual motion.

[0038] S10212, Feature Alignment. Due to the differences in hardware sampling frequencies among various sensors, cubic spline interpolation is used to resample the signals of each physical channel based on a preset time resolution, thereby completing the feature alignment of the time axis. This step effectively eliminates the temporal misalignment error of multimodal data, ensuring that the data points of all modes strictly correspond at the same timestamp.

[0039] S10213. Extract multi-dimensional temporal variation features. Within a set sliding time window, calculate the features of each modal signal sequentially. Time-domain statistical features may include mean, variance, and peak-to-peak value, while frequency-domain features cover the energy proportion of each frequency band, etc. Finally, these features are concatenated to form a high-dimensional feature vector sequence, thus extracting multi-dimensional temporal variation features.

[0040] S1022. Input the temporal variation features into a sleep state recognition model pre-deployed on the local edge and built based on a lightweight deep learning network. The sleep stage state is calculated in real time through the temporal convolutional layer and classification fully connected layer inside the model.

[0041] In one optional implementation, the sleep stages include at least three states: initial sleep, light sleep, and deep sleep. This sleep state recognition model compresses the network size using knowledge distillation (utilizing a large-parameter teacher model to guide a lightweight student model) or weight quantization techniques to adapt to local edge deployments, reducing computational latency and power consumption while maintaining inference accuracy. The model structure includes a temporal convolutional layer that processes the input temporally changing feature sequence using a one-dimensional dilated causal convolution operator. The causal convolution ensures that the model strictly follows the temporal order during inference without leaking future information, while the dilation mechanism effectively increases the receptive field in the temporal dimension, capturing the long-term evolutionary dependencies of physiological signals over time with lower computational cost. The extracted high-level abstract features are then passed to a classification fully connected layer, where a Softmax activation function is used for probability mapping. The output determines the maximum probability classification label for the current sliding time window, indicating whether it belongs to the initial sleep, light sleep, or deep sleep state, thus obtaining the sleep stage state in real time.

[0042] S1023. Parallel monitoring of body movement frequency in body movement data, heart rate increase in heart rate data, and blood oxygen decrease in blood oxygen data. When the body movement frequency, heart rate increase, or blood oxygen decrease exceeds the pre-set safety baseline for the individual, it is determined to be in an abnormal fluctuation state.

[0043] A monitoring thread is established in the background to maintain the normal physiological baseline for recent time windows. If the calculated body movement frequency within the current time window exceeds the upper limit of the baseline tolerance (e.g., violent turning over within 30 consecutive seconds), or if there is a sudden jump in the increase in heart rate (e.g., a sudden increase of 20% from the resting baseline), or if the decrease in blood oxygen causes the absolute value to drop to a trough below the preset threshold (e.g., below 92%), an interruption signal is generated, indicating that the body has deviated from the stable resting trajectory and is in an abnormal fluctuation state.

[0044] S103, based on the sleep stage state, generates a combined sleep aid parameter configuration including an initial threshold through a preset parameter mapping rule. The combined sleep aid parameter configuration is used to control the operation of the sleep aid device.

[0045] In one optional implementation, the combined sleep-aid parameter configuration includes at least audio stimulation parameters, light stimulation parameters, and magnetic stimulation parameters. The electronic device stores a multi-dimensional parameter mapping rule matrix to define the physical quantity conversion relationships between different sleep stages and the audio, light, and magnetic stimulation parameters.

[0046] S1031. When the sleep stage is in the state of falling asleep, generate audio stimulation parameters including low-frequency sleep-aiding frequency, fixed slow-changing rhythm and decreasing volume, as well as light stimulation parameters including gradually decreasing low brightness and sleep-aiding warm color temperature.

[0047] When the device is in a sleep-inducing state, corresponding audio stimulation parameters are generated according to parameter mapping rules: sound waves containing binaural beats or pink noise characteristics are configured. Pink noise effectively masks sudden background noise in the environment, while binaural beats use the slight frequency difference between the left and right ears to set the frequency in the slow-wave band of the brainwave to assist the brain in following the rhythm. The beat is set to a fixed, slowly varying rhythm simulating steady breathing, and the output volume is set to decrease smoothly from an initial safe decibel value (e.g., 40 decibels) according to an exponential curve to avoid the risk of awakening from a sudden drop in volume. Simultaneously, corresponding light stimulation parameters are generated: the spectrum is configured to strictly avoid short-wave blue light (wavelength approximately 400-480 nanometers) to promote sleep with a warm color temperature, maintaining normal melatonin secretion, and the light flux is gradually reduced from a low initial value according to a preset slow decrease slope until the device is turned off. The initial values ​​assigned to each parameter at this point are the initial thresholds.

[0048] S1032. When the sleep stage is light sleep or deep sleep, generate magnetic stimulation parameters including magnetic stimulation intensity, magnetic stimulation frequency and single duration within the preset brain nerve safety threshold range, and gradually decrease the playback volume of the audio stimulation parameters as the sleep stage progresses from light sleep to deep sleep until playback stops completely.

[0049] Once the user enters a light or deep sleep state, the playback volume of the audio stimulation parameters is gradually reduced until it is muted, and the light intervention is simultaneously turned off to maintain a pure dark environment. At this time, the parameter mapping rules trigger transcranial magnetic field intervention or weak alternating magnetic field stimulation. Within the preset brain nerve safety threshold range, a magnetic stimulation intensity at the microtesla level (the magnetic stimulation intensity is limited within the preset device output safety boundary to reduce the risk of discomfort during home use), a magnetic stimulation frequency in sync with slow brain waves, and magnetic stimulation parameters consisting of a pulse train and a single duration are generated.

[0050] In one optional implementation, based on the sleep stage state and through preset parameter mapping rules, a joint sleep-aid parameter configuration including an initial threshold is generated, which further includes: S1033. Extract sleep slow wave signals in real time from continuously collected EEG data, and use phase-locked loop algorithm to calculate the real-time phase sequence of sleep slow wave signals.

[0051] The biological marker of deep sleep is high-amplitude sleep slow waves (mainly distributed in the 0.5 Hz to 2 Hz frequency band). Within the time domain, EEG data is subjected to specific narrow-band digital filtering to extract slow waves in the target frequency band. Subsequently, a digital phase-locked loop (PLL) algorithm is introduced for phase tracking: the phase error between the input sleep slow wave signal and the local reference signal is compared by a phase detector, and the error is smoothed by a low-pass loop filter. Frequency and phase are corrected in real time, and a precise real-time phase sequence is continuously output.

[0052] S1034. Using the real-time phase sequence as the closed-loop feedback trigger condition, when the sleep slow wave signal is detected to reach a specific upward phase window, the stimulation pulse output corresponding to the joint sleep aid parameter configuration is precisely and synchronously triggered to achieve positive enhancement modulation of the sleep slow wave amplitude.

[0053] When the phase angle of the slow wave is detected to enter a specific ascending phase window, the hardware blockade is released, and the sleep aid device is commanded to precisely and synchronously output short auditory or magnetic stimulation pulses. At this time, the intervention pulse matches the rhythm of the brain's spontaneous activity (the ascending phase usually corresponds to the depolarization state of cortical neurons), achieving positive modulation of the amplitude of slow wave sleep, thereby deepening sleep and aiding memory consolidation.

[0054] S104 When an abnormal fluctuation is detected, the parameter conflict suppression mechanism is triggered to safely adjust the combined sleep aid parameter configuration, and the sleep aid device is controlled to work according to the adjusted combined sleep aid parameter configuration.

[0055] In one alternative implementation, when an abnormal fluctuation is detected, a parameter conflict suppression mechanism is triggered to safely adjust the combined sleep aid parameter configuration, including the following steps: S1041. Analyze the specific abnormal type of the abnormal fluctuation state. By tracing the abnormal flag bit thrown in S102, identify whether the current abnormality is induced by the blood oxygenation domain, heart rate domain, or body motion domain.

[0056] S1042. When the specific abnormality type is an abnormal drop in blood oxygen, it is determined that there is a physiological safety conflict in the configuration of the combined sleep aid parameters. The output of audio stimulation parameters, light stimulation parameters and magnetic stimulation parameters is immediately and forcibly suspended, and health reminder parameters for triggering external alarms are generated.

[0057] An abnormal drop in blood oxygen levels (such as that seen in sleep apnea syndrome) indicates a state of physiological inhibition. To restore upper airway patency, the parameter conflict suppression mechanism sends a zeroing command to the drive interface, suspending all hardware output channels. Simultaneously, health reminder parameters are generated (e.g., driving the vibration motor of the wearable device to provide localized tactile feedback, or triggering a gentle alarm from an external device), triggering status reminders to alert the user, monitoring terminal, or associated devices to the current sleep state. It can also prompt the user to adjust their sleeping position (e.g., from supine to side-lying) to restore breathing patency.

[0058] S1043. When the specific abnormality type is frequent body movement or abnormally high heart rate, the output intensity of the magnetic stimulation parameters is dynamically reduced according to the preset attenuation ratio, and the audio stimulation parameters are switched to soothing white noise with calming and calming effects to inhibit the deterioration of the state and help to fall back asleep.

[0059] At this time, the body may be on the edge of micro-awakening. The output intensity of the magnetic stimulation parameters is dynamically reduced by multiplying the attenuation ratio coefficient. At the same time, the audio stimulation parameters are switched and soothing white noise is loaded to form an acoustic barrier, which isolates the environmental background noise interference that may cause micro-awakening (such as distant vehicle sounds or sudden abnormal noises) and inhibits the transformation of the micro-awakening state into full awakening.

[0060] S105: Based on sleep evaluation data, calculate the sleep improvement score using a preset periodic evaluation model, and update the initial threshold of the combined sleep aid parameter configuration based on the sleep improvement score.

[0061] In one optional implementation, a sleep improvement score is calculated based on sleep evaluation data using a pre-defined periodic evaluation model, including: S1051. Input the sleep index statistics and subjective feedback data into the periodic evaluation model constructed based on the multi-dimensional weighted evaluation algorithm, and calculate the objective physiological improvement score and the subjective psychological satisfaction score respectively.

[0062] The periodic evaluation model first uses the range transformation method to standardize and map the dimensional data. Specifically, it performs inverse standardization on negative indicators (such as sleep latency duration and total number of nighttime awakenings, with higher values ​​indicating poorer sleep) and positive standardization on positive indicators (such as the percentage of deep sleep and sleep continuity index). The sum of these values ​​yields an objective physiological improvement score. Similarly, the questionnaire scores are quantified and mapped to obtain a subjective psychological satisfaction score.

[0063] S1052. A weighted fusion calculation is performed based on the weight coefficients pre-assigned to the objective physiological improvement score and the subjective psychological satisfaction score to obtain a sleep improvement score that characterizes the overall sleep quality improvement effect.

[0064] Sleep improvement score The weighted fusion calculation formula is as follows: In the formula, This represents the calculated sleep improvement score; This represents the objective physiological improvement score obtained after standardizing the statistical data of objective sleep indicators; This represents the subjective psychological satisfaction score obtained after quantifying subjective feedback data. This represents the weighting coefficients pre-assigned to the objective physiological improvement score; This represents the weighting coefficient pre-assigned to the subjective psychological satisfaction score, and satisfies... .

[0065] Through the above-mentioned integrated calculation, it is possible to simultaneously combine changes in the body's objective physiological indicators with the user's subjective self-awareness, thus avoiding the bias caused by a single-dimensional evaluation.

[0066] In one alternative implementation, the initial thresholds of the combined sleep aid parameter configuration are updated based on the sleep improvement score, including: S1053. Determine whether the sleep improvement score is lower than the preset improvement expectation threshold.

[0067] S1054. If the sleep improvement score is lower than the expected improvement threshold, a weakness diagnosis is performed on the sleep indicator statistics to identify the target sleep indicator with the lowest score. By tracing back the historical distribution range of each input sub-score, the indicator with the worst score (e.g., an indicator that is consistently lower than the group average benchmark) is intercepted as the target that urgently needs to be improved.

[0068] S1055. Based on the target sleep indicators and combined with the individual sleep-wake patterns of the user, the initial thresholds of the intervention time interval, stimulation intensity change rate, and trigger time baseline value in the combined sleep aid parameter configuration are recalculated and dynamically updated using a preset circadian rhythm mathematical model to generate a revised intervention plan.

[0069] Internally deployed is a circadian rhythm mathematical model. By inputting the user's sleep and wake times over multiple consecutive days, it calculates the inherent cycle and natural sleep phase attributes of the individual's biological clock. If the weakness is a "low sleep latency score," the model infers a shift in the natural sleep phase rhythm (i.e., a tendency to sleep late). In the initial threshold update of the next cycle, the audio playback time interval (intervention time interval) is dynamically delayed to better align with the actual sleep onset time, and the rate of light dimming (stimulus intensity change rate) is reduced to provide the body with a more sufficient preparation period for melatonin secretion. If the weakness is "insufficient deep sleep," the trigger time of magnetic stimulation (trigger time baseline value) is adjusted during the transition from light sleep to deep sleep to prolong the deep sleep consolidation period. Thus, a data-driven personalized closed loop is formed over a long period, completing the adaptive update of individual intervention parameters.

[0070] This invention also provides an adaptive adjustment device based on sleep-aid parameters, referring to... Figure 3 The diagram shows a functional block diagram of an adaptive adjustment device 300 based on sleep-aid parameters according to the present invention. The device may include the following modules: The data acquisition module 301 is used to collect the user's multimodal physiological characteristic data and obtain the user's sleep evaluation data within a preset period; The state recognition module 302 is used to input multimodal physiological feature data into a preset sleep state recognition model to obtain sleep stage state and abnormal fluctuation state. The parameter generation module 303 is used to generate a combined sleep aid parameter configuration containing an initial threshold according to the sleep stage state and through a preset parameter mapping rule. The combined sleep aid parameter configuration is used to control the operation of the sleep aid device. The safety suppression module 304 is used to trigger the parameter conflict suppression mechanism to safely adjust the combined sleep aid parameter configuration when an abnormal fluctuation state is detected, and to control the sleep aid device to work according to the adjusted combined sleep aid parameter configuration. The evaluation update module 305 is used to calculate the sleep improvement score based on the sleep evaluation data using a preset periodic evaluation model, and update the initial threshold of the combined sleep aid parameter configuration based on the sleep improvement score.

[0071] Based on the same inventive concept, another embodiment of the present invention provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus. Memory, used to store computer programs; When a processor executes a program stored in memory, it implements the adaptive adjustment method based on sleep-aid parameters of the present invention.

[0072] The communication bus mentioned in the aforementioned electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the figure, but this does not mean that there is only one bus or one type of bus. The communication interface is used for communication between the aforementioned terminal and other devices. The memory can include RAM, or it can include NVM, such as at least one disk storage device. Optionally, the memory can also be at least one storage device located remotely from the aforementioned processor.

[0073] The processors mentioned above can be general-purpose processors, including CPUs, network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0074] Furthermore, to achieve the above objectives, embodiments of the present invention also propose a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the adaptive adjustment method based on sleep-aid parameters of the embodiments of the present invention.

[0075] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, embodiments of the present invention can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of computer program products embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0076] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, electronic devices, apparatuses, and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0077] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0078] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0079] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. "And / or" indicates that either one or both can be chosen. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes the element.

[0080] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An adaptive adjustment method based on sleep-aiding parameters, characterized in that, include: Collect multimodal physiological characteristic data of users and obtain sleep evaluation data of the users within a preset period; The multimodal physiological feature data is input into a preset sleep state recognition model to obtain sleep stage states and abnormal fluctuation states. Based on the sleep stage state, a combined sleep aid parameter configuration including an initial threshold is generated through a preset parameter mapping rule. The combined sleep aid parameter configuration is used to control the operation of the sleep aid device. When the abnormal fluctuation state is detected, the parameter conflict suppression mechanism is triggered to safely adjust the combined sleep aid parameter configuration, and the sleep aid device is controlled to work according to the adjusted combined sleep aid parameter configuration; Based on the sleep evaluation data, a sleep improvement score is calculated using a preset periodic evaluation model, and the initial threshold of the combined sleep aid parameters is updated based on the sleep improvement score.

2. The adaptive adjustment method based on sleep-aid parameters as described in claim 1, characterized in that, The multimodal physiological data includes at least body movement data, heart rate data, respiratory data, electroencephalogram (EEG) data, and blood oxygen saturation data. The step of inputting the multimodal physiological feature data into a preset sleep state recognition model to obtain sleep stage states and abnormal fluctuation states includes: Signal denoising and feature alignment are performed on the body motion data, heart rate data, respiratory data, and blood oxygen data, and multi-dimensional temporal variation features are extracted. The temporal variation features are input into the sleep state recognition model, which is pre-deployed on the local edge and built based on a lightweight deep learning network. The sleep stage state is calculated in real time through the temporal convolutional layer and classification fully connected layer inside the model. The system monitors the body movement frequency in the body movement data, the heart rate increase in the heart rate data, and the blood oxygen decrease in the blood oxygen data in parallel. When the body movement frequency, the heart rate increase, or the blood oxygen decrease exceeds a pre-set safety baseline for the individual, it is determined to be in the abnormal fluctuation state.

3. The adaptive adjustment method based on sleep-aid parameters as described in claim 1, characterized in that, The sleep stages include at least the initial sleep state, light sleep state, and deep sleep state; the combined sleep-aid parameter configuration includes at least audio stimulation parameters, light stimulation parameters, and magnetic stimulation parameters; The step of generating a combined sleep-aid parameter configuration including an initial threshold based on the sleep stage state and through a preset parameter mapping rule includes: When the sleep stage state is the state of falling asleep, the audio stimulation parameters, which include a low-frequency sleep-aiding frequency, a fixed slow-changing rhythm and a decreasing volume, and the light stimulation parameters, which include a gradually decreasing low brightness and a sleep-aiding warm color temperature, are generated. When the sleep stage is in the light sleep state or the deep sleep state, the magnetic stimulation parameters, including magnetic stimulation intensity, magnetic stimulation frequency and single duration, are generated within a preset brain nerve safety threshold range. The playback volume of the audio stimulation parameters is gradually reduced from the light sleep state to the deep sleep state until playback stops completely.

4. The adaptive adjustment method based on sleep-aid parameters as described in claim 3, characterized in that, When the abnormal fluctuation state is detected, the parameter conflict suppression mechanism is triggered to safely adjust the combined sleep aid parameter configuration, including: Analyze the specific anomaly type of the abnormal fluctuation state; When the specific abnormality type is an abnormal drop in blood oxygen, it is determined that there is a physiological safety conflict in the configuration of the combined sleep aid parameters. The output of the audio stimulation parameters, the light stimulation parameters, and the magnetic stimulation parameters is immediately and forcibly suspended, and health reminder parameters for triggering external alarms are generated. When the specific abnormality type is a state of frequent body movement or an abnormally high heart rate, the output intensity of the magnetic stimulation parameter is dynamically reduced according to a preset attenuation ratio, and the audio stimulation parameter is switched to soothing white noise with calming and calming effects to inhibit the deterioration of the state and help to fall back asleep.

5. The adaptive adjustment method based on sleep-aid parameters as described in claim 1, characterized in that, The sleep evaluation data includes subjective feedback data and sleep index statistics extracted from sleep monitoring records. The sleep index statistics include sleep latency duration, total number of nighttime awakenings, percentage of deep sleep, and sleep continuity index. The step of calculating a sleep improvement score based on the sleep evaluation data using a preset periodic evaluation model includes: The sleep index statistics and the subjective feedback data are input into the periodic evaluation model constructed based on the multi-dimensional weighted evaluation algorithm to calculate the objective physiological improvement score and the subjective psychological satisfaction score, respectively. The sleep improvement score, which characterizes the overall sleep quality improvement effect, is obtained by weighted fusion calculation based on the weight coefficients pre-assigned to the objective physiological improvement score and the subjective psychological satisfaction score.

6. The adaptive adjustment method based on sleep-aid parameters as described in claim 5, characterized in that, The step of updating the initial threshold configured for the combined sleep aid parameters based on the sleep improvement score includes: Determine whether the sleep improvement score is lower than a preset expected improvement threshold; If the sleep improvement score is lower than the expected improvement threshold, then the sleep index statistics are used to diagnose weaknesses and locate the target sleep index with the lowest score. Based on the target sleep index and combined with the individual user's sleep-wake pattern, the initial thresholds of the intervention time interval, stimulation intensity change rate, and trigger time benchmark value in the combined sleep aid parameter configuration are recalculated and dynamically updated using a preset circadian rhythm mathematical model to generate a revised intervention plan.

7. The adaptive adjustment method based on sleep-aid parameters as described in claim 2, characterized in that, The step of generating a combined sleep-aid parameter configuration including an initial threshold based on the sleep stage state and through a preset parameter mapping rule also includes: Sleep slow-wave signals are extracted in real time from the continuously collected EEG data, and the real-time phase sequence of the sleep slow-wave signals is calculated using a phase-locked loop algorithm. Using the real-time phase timing as a closed-loop feedback trigger condition, when the sleep slow wave signal is detected to have reached a specific upward phase window, a stimulation pulse output corresponding to the joint sleep aid parameter configuration is precisely and synchronously triggered to achieve positive enhancement modulation of the sleep slow wave amplitude.

8. An adaptive adjustment device based on sleep-aid parameters, characterized in that, include: The data acquisition module is used to collect the user's multimodal physiological characteristic data and obtain the user's sleep evaluation data within a preset period; The state recognition module is used to input the multimodal physiological feature data into a preset sleep state recognition model to obtain sleep stage states and abnormal fluctuation states. The parameter generation module is used to generate a combined sleep aid parameter configuration containing an initial threshold according to the sleep stage state and through a preset parameter mapping rule. The combined sleep aid parameter configuration is used to control the operation of the sleep aid device. The safety suppression module is used to trigger the parameter conflict suppression mechanism to safely adjust the combined sleep aid parameter configuration when the abnormal fluctuation state is detected, and to control the sleep aid device to work according to the adjusted combined sleep aid parameter configuration. The evaluation update module is used to calculate a sleep improvement score based on the sleep evaluation data using a preset periodic evaluation model, and to update the initial threshold configured for the combined sleep aid parameters based on the sleep improvement score.

9. An electronic device, characterized in that, include: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, and when the computer program is executed by the at least one processor, it implements the adaptive adjustment method based on sleep aid parameters as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the adaptive adjustment method based on sleep-aid parameters as described in any one of claims 1 to 7.

Images