Wearable monitoring system for sleep-wake rhythm disorder and method thereof

Abstract

The invention relates to the field of sleep medicine, in particular to a wearable monitoring system for sleep-awakening rhythm disorder and a method thereof. Comprising a monitoring module, a signal processing module, a signal analysis module and an intervention scheme execution module. The monitoring module is used for acquiring collective acceleration, PPG signals and emotional state data; the signal processing module is used for filtering and down-sampling; the signal analysis module extracts time-domain and non-time-domain features by using a pulse neural network; and the intervention scheme execution module selects a personalized phototherapy and vibration intervention scheme by adopting a Q-learning algorithm according to the feature data. According to the system, through technical innovation and ingenious design, high intelligence and individuation are achieved, the sleep quality and the daytime function of the user are effectively improved, and a new tool and method are provided for sleep medical research.

Description

Technical Field

[0001] The present invention relates to the field of sleep medicine, particularly to a wearable monitoring system and method for sleep - wake rhythm disorders. Background Art

[0002] Sleep - wake rhythm disorder is a common sleep problem that affects the quality of life and health of millions of people worldwide. With the accelerating pace of modern life and the popularity of electronic devices, this problem has become increasingly serious. In recent years, the development of wearable device technology has provided new possibilities for monitoring and improving sleep - wake rhythms.

[0003] Currently, there are various sleep monitoring devices on the market, which mainly collect users' activity and physiological data through accelerometers and photoplethysmography (PPG) sensors. These devices can usually record the user's sleep time, sleep stages, and basic physiological indicators such as heart rate and respiratory rate. However, the existing technologies still have obvious deficiencies in the following aspects:

[0004] Firstly, most devices use a fixed sampling frequency, which results in inconsistent data quality in different activity states, either sacrificing battery life or reducing data accuracy at critical moments. Secondly, the existing signal processing algorithms mainly focus on time - domain or frequency - domain analysis, making it difficult to comprehensively capture the complex dynamic characteristics of sleep - wake rhythms. Thirdly, although some high - end devices provide simple sleep suggestions, these suggestions are often based on general rules and lack the ability of personalization and dynamic adjustment.

[0005] More critically, the existing technologies have limited capabilities in integrating multi - source data, extracting deep - level features, and providing precise interventions. They usually regard sleep monitoring and intervention as two independent processes, lacking a closed - loop feedback mechanism to continuously optimize intervention strategies. In addition, most systems do not consider the impact of emotional states on sleep - wake rhythms, which may lead to poor intervention effects in some cases.

[0006] Finally, the existing technologies also face challenges in terms of user - friendliness and long - term compliance. Complicated operations and lack of intuitive data display often cause users to gradually abandon the use after the initial enthusiasm, affecting the long - term effectiveness of the devices.

[0007] In view of the above problems, there is an urgent need for a system that can comprehensively and accurately monitor sleep - wake rhythms and provide personalized and adaptive intervention solutions. Such a system should be able to intelligently process multi - source data, deeply analyze sleep - wake patterns, and optimize intervention strategies through continuous learning, while maintaining a good user experience. Summary of the Invention

[0008] The present invention aims to solve the above technical problems and provides an innovative wearable monitoring system and method for sleep-wake rhythm disorders. By integrating advanced sensing technologies, intelligent signal processing algorithms, adaptive neural networks, reinforcement learning and other technologies, the system realizes the full-process intelligent management from data collection to personalized intervention.

[0009] The present invention proposes a wearable monitoring system for sleep-wake rhythm disorders, including:

[0010] A monitoring module, configured to:

[0011] Collect the body movement acceleration information and PPG signals of the user;

[0012] Detect the emotional state data of the user;

[0013] A signal processing module, electrically connected to the monitoring module, configured to:

[0014] Receive the body movement acceleration information, PPG signals and emotional state data sent by the monitoring module;

[0015] Perform filtering and downsampling processing on the body movement acceleration information, PPG signals and emotional state data;

[0016] A signal analysis module, electrically connected to the signal processing module, configured to:

[0017] Receive the processed body movement acceleration information and PPG signals sent by the signal processing module;

[0018] Analyze the processed body movement acceleration information and PPG signals by using a spiking neural network to obtain time-domain and non-time-domain features;

[0019] An intervention plan execution module, electrically connected to the signal analysis module, configured to:

[0020] Receive the time-domain and non-time-domain features sent by the signal analysis module;

[0021] According to the phase deviation, nighttime sleep wake-up and sleep efficiency data in the time-domain and non-time-domain features, select a personalized intervention plan by using a Q-learning-based reinforcement learning algorithm;

[0022] Execute the personalized intervention plan, where the personalized intervention plan includes light therapy and vibration.

[0023] Preferably, the signal analysis module includes:

[0024] A PPG signal time series feature extraction unit, configured to:

[0025] Calculate the phase deviation of the PPG signal;

[0026] Provide prior knowledge for individual nighttime sleep efficiency assessment;

[0027] Obtain the accuracy of phase correction, sleep efficiency, and the number of awakenings as a reward function based on the nighttime sleep assessment results.

[0028] Preferably, the spiking neural network is an adaptive hybrid spiking neural network, including:

[0029] A hybrid spiking model based on the time domain, used for:

[0030] Statistically analyze the time characteristics of the time series using the method of spike counting;

[0031] Provide a method for quantifying the difference in spike occurrence time;

[0032] A scale-normalized adaptive spiking model, used for:

[0033] Overcome the linear relationship of the time series;

[0034] Dynamically adjust the activation threshold of neurons according to the training process;

[0035] Adjust the pulse width according to the standard deviation and sample entropy of the input time series;

[0036] A periodic spiking model based on the time-frequency domain, used for:

[0037] Measure the pulse time interval of the time series through Fourier transform and time-related entropy;

[0038] Adopt a scale-normalized periodic spiking model.

[0039] Preferably, the system adopts an adaptive sampling technique, used for:

[0040] Increase the sampling interval when the body movement data is in the low amplitude segment;

[0041] Reduce the sampling interval when the body movement data is in the high amplitude segment;

[0042] Optimize the sampling interval to 30% of the original sampling interval;

[0043] Apply the adaptive sampling technique to PPG data acquisition.

[0044] Preferably, the intervention plan execution module includes:

[0045] A light therapy execution unit, used for:

[0046] Adjust the light intensity and duration according to the personalized intervention plan;

[0047] Execute light therapy intervention within a preset time period;

[0048] A vibration execution unit, configured to:

[0049] Adjust the vibration intensity and frequency according to the personalized intervention plan;

[0050] Execute vibration intervention within a preset time;

[0051] An intervention effect evaluation unit, configured to:

[0052] Collect the PPG signal and body movement data after the intervention;

[0053] Evaluate the intervention effect and feedback it to the intervention plan execution module for strategy optimization.

[0054] Preferably, it further includes:

[0055] A data storage module, electrically connected to the signal processing module and the signal analysis module, and configured to:

[0056] Store the original physiological data and processed feature data of the user;

[0057] Record the intervention history and effect evaluation results of the user;

[0058] A user interface module, electrically connected to the intervention plan execution module, and configured to:

[0059] Display the sleep-wake rhythm state and intervention suggestions of the user;

[0060] Receive the feedback and personalized settings of the user.

[0061] Preferably, the intervention plan execution module further includes:

[0062] A circadian rhythm phase response curve modeling unit, configured to:

[0063] Establish an individualized circadian rhythm phase response curve based on the PPG feature data of the user; provide a theoretical basis for the formulation of the personalized intervention plan.

[0064] Preferably, the system further includes:

[0065] A mobile terminal, wirelessly connected to the monitoring module, and configured to:

[0066] Receive and store the data collected by the monitoring module;

[0067] Execute data processing and analysis algorithms;

[0068] Display the analysis results and intervention suggestions.

[0069] Preferably, it is characterized in that:

[0070] The mobile terminal includes applications based on the Android operating system for:

[0071] Visually displaying the user's sleep-wake rhythm data;

[0072] Providing a personalized intervention plan setting interface;

[0073] Recording and analyzing the long-term sleep-wake rhythm change trend of the user.

[0074] A wearable monitoring method for sleep-wake rhythm disorder, using the described system, includes the following steps:

[0075] S1: Collect the user's body movement acceleration information, PPG signal, and emotional state data through the monitoring module;

[0076] S2: The signal processing module filters and down-samples the collected data, and the filtered data is input into the signal analysis module;

[0077] S3: The signal analysis module uses a pulsed neural network to analyze the processed body movement acceleration information and PPG signal to obtain the phase and circadian rhythm deviation of the individual PPG signal;

[0078] S4: Use a reinforcement learning algorithm based on Q-learning to optimize the individual's intervention strategy, and the optimization of the intervention strategy is achieved through the following steps:

[0079] Input the individual classification features into the intervention module to select personalized strategies for light therapy or vibration;

[0080] Combine the time series and event analysis module to evaluate night awakenings;

[0081] Combine the reinforcement learning algorithm to analyze the individual's sleep efficiency and wake-up times;

[0082] Calculate the accuracy of phase deviation correction as the input of the reward function;

[0083] S5: Execute the personalized intervention plan, including light therapy and vibration intervention;

[0084] S6: Evaluate the intervention effect and update the intervention strategy.

[0085] The wearable monitoring system and method for sleep-wake rhythm disorder of the present invention demonstrate significant advantages and innovations at multiple levels. From a macroscopic perspective, the system constructs a complete closed-loop monitoring-analysis-intervention framework, realizing the comprehensive intelligence of sleep health management. This holistic solution not only improves the accuracy of sleep disorder diagnosis and intervention but also provides rich data support for research in related fields.

[0086] At the data acquisition level, the adaptive sampling technique adopted by the present invention ingeniously solves the limitations of the traditional fixed sampling rate method. By dynamically adjusting the sampling frequency, the system can significantly extend the battery life of the device while ensuring the accuracy of key data. This not only improves the user experience but also ensures the feasibility of long-term continuous monitoring, laying a foundation for capturing the long-term change trend of the sleep-wake rhythm.

[0087] In terms of signal processing and feature extraction, the adaptive hybrid pulse neural network of the present invention demonstrates powerful performance. By integrating time-domain, frequency-domain, and nonlinear analysis techniques, this network can extract richer and more accurate features from complex physiological signals. This multi-dimensional feature extraction method greatly improves the system's ability to identify sleep-wake patterns, providing a reliable data basis for subsequent personalized interventions.

[0088] In the formulation and implementation of intervention strategies, the reinforcement learning algorithm based on Q-learning of the present invention achieves true personalization and adaptability. The system can continuously adjust and optimize the intervention plan according to the user's real-time feedback. This dynamic learning ability enables the intervention effect to continuously improve over time. This not only improves the effectiveness of the intervention but also adapts to the long-term changes in the user's lifestyle and physiological state.

[0089] The system of the present invention also demonstrates unique advantages in the coordination between modules. A tight information flow and feedback loop are formed among the monitoring module, signal processing module, analysis module, and intervention execution module. This highly integrated design not only improves the overall efficiency of the system but also enables the complementary and collaborative enhancement of the functions of each module. For example, the real-time feedback of the intervention effect can be used to optimize the parameters of the signal processing algorithm, thereby further improving the accuracy of feature extraction.

[0090] In solving technical contradictions, the present invention also shows innovation. For example, the system successfully balances the contradiction between the need for real-time monitoring and device energy consumption through an intelligent scheduling algorithm. At the same time, by using edge computing technology, the system realizes the local execution of complex algorithms while ensuring data privacy, solving the conflict between data security and processing efficiency.

[0091] From a microscopic perspective, the present invention has breakthroughs in many technical details. For example, the adaptive filtering algorithm adopted in PPG signal processing can effectively remove various complex interference signals, significantly improving the accuracy of heart rate variability analysis. Another example is that the integrated emotion state detection function of the system adds a new dimension to sleep-wake rhythm analysis, enabling intervention strategies to more comprehensively consider the physical and mental state of the user.

[0092] Generally speaking, through multiple technological innovations and ingenious system designs, the present invention has successfully constructed a highly intelligent, personalized, and user-friendly sleep-wake rhythm monitoring and intervention system. This comprehensive solution can not only effectively improve the user's sleep quality and daytime function but also provide new tools and methods for sleep medicine research. With the further development of technology and the accumulation of clinical data, this system is expected to play an increasingly important role in the field of sleep health management and make important contributions to improving the global sleep health situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] Figure 1 It is the overall logic block diagram of the system of the present invention;

[0094] Figure 2 It is the logic block diagram of the signal analysis module of the present invention;

[0095] Figure 3 It is the logic block diagram of the intervention plan execution module of the present invention;

[0096] Figure 4 It is the application program logic block diagram of the mobile terminal of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0097] Please refer to the attached Figures 1-4 , the present invention provides a wearable monitoring system and method for sleep-wake rhythm disorders. The system aims to real-time monitor the user's physiological data through an intelligent wristband, analyze the sleep-wake rhythm pattern, and provide personalized intervention strategies to improve sleep quality and daytime function.

[0098] The wearable monitoring system of the present invention includes a monitoring module 1, a signal processing module 2, a signal analysis module 3, and an intervention plan execution module 4. These modules work together to form a closed-loop monitoring-analysis-intervention system.

[0099] The monitoring module 1 is mainly responsible for collecting the user's body movement acceleration information and PPG (photoplethysmography) signals, and can also detect the user's emotional state data. Preferably, the monitoring module 1 uses high-precision triaxial acceleration sensors and photoelectric sensors, which can collect body movement data at a relatively high sampling rate (such as 100Hz) and collect PPG signals at an appropriate sampling rate (such as 25Hz). In addition, in an embodiment of the present invention, the emotional state data can be indirectly obtained by analyzing the user's heart rate variability (HRV) and skin electrical activity (EDA).

[0100] The signal processing module 2 is electrically connected to the monitoring module 1 and is used to receive and process the raw data. This module uses digital filtering technology to preprocess the signals, removing high-frequency noise and baseline drift. For example, for body movement acceleration signals, a high-pass filter with a cut-off frequency of 0.5 Hz can be used to remove the gravity component, and for PPG signals, a band-pass filter with a frequency range of 0.5 - 5 Hz can be used to retain the effective signals. In addition, the signal processing module 2 also performs downsampling operations, reducing the sampling rate of the body movement signals to 20 Hz and the PPG signals to 10 Hz to reduce the computational burden of subsequent processing.

[0101] The signal analysis module 3 is the core component of this system. It uses innovative pulse neural network technology to analyze the processed body movement and PPG signals and extract time-domain and non-time-domain features. The main innovation of this module lies in the adoption of an adaptive hybrid pulse neural network, which can effectively capture the non-linear spatio-temporal relationships of time series data.

[0102] Specifically, the adaptive hybrid pulse neural network contains three pulse models: a time-domain-based hybrid pulse model, a scale-normalized adaptive pulse model, and a time-frequency-domain-based periodic pulse model. This multi-model fusion design enables the network to adapt to different types of time series features, improving the comprehensiveness and accuracy of feature extraction.

[0103] The time-domain-based hybrid pulse model is mainly used to statistically analyze the time features of time series. This model uses a pulse counting method, providing an effective way to quantify the time differences of pulse occurrences. In practical applications, a time window (such as 10 seconds) can be set to count the number and time distribution of pulses within the window, thereby capturing short-term time series patterns.

[0104] The design purpose of the scale-normalized adaptive pulse model is to overcome the linear relationship of time series. This model has two key adaptive features: neuron activation threshold adaptation and pulse width adaptation. The activation threshold of the neuron is dynamically adjusted during the training process. Initially, it can be set to 1.5 times the mean of the input signal, and then gradually adjusted according to the output error of the network. The pulse width is adjusted according to the standard deviation and sample entropy of the input time series. The larger the standard deviation, the wider the pulse width to adapt to the degree of signal variation.

[0105] The time-frequency-domain-based periodic pulse model measures the pulse time interval of time series through Fourier transform and time-related entropy. This model uses a scale-normalized periodic pulse model, which can effectively capture the periodic features of signals. In practical applications, an appropriate time window (such as 30 minutes) can be selected for Fourier transform to extract the main frequency components and calculate the time-related entropy to quantify the regularity of the signal.

[0106] The outputs of these three pulse models are weighted and fused to form the final feature representation. The weights can be dynamically adjusted through the backpropagation algorithm. Initially, they can all be set to 1 / 3 and then gradually optimized according to the performance of each model.

[0107] The signal analysis module 3 also includes a PPG signal time series feature extraction unit 31, which is specifically used to analyze the features of the PPG signal. It mainly completes three tasks: calculating the phase deviation of the PPG signal, providing prior knowledge for the evaluation of individual nocturnal sleep efficiency, and obtaining the accuracy of phase correction, sleep efficiency, and the number of awakenings as a reward function based on the nocturnal sleep evaluation results.

[0108] The phase deviation is calculated using the following method: First, the instantaneous phase of the PPG signal is obtained through the Hilbert transform and then compared with the ideal 24-hour periodic signal to calculate the deviation. The mathematical expression is as follows:

[0109]

[0110] Δφ(t)=φ(t)-2πt / 24,

[0111] where φ(t) is the instantaneous phase of the PPG signal, H[s(t)] is the Hilbert transform of the signal s(t), and Δφ(t) is the phase deviation.

[0112] The following indicators are used to evaluate sleep efficiency: total sleep time (TST), sleep latency (SL), and sleep efficiency (SE). These indicators can be estimated by analyzing the fluctuation characteristics of the PPG signal and body movement information. For example, sleep efficiency can be calculated using the following formula:

[0113]

[0114] where TIB (Time in Bed) is the time in bed.

[0115] These features and evaluation results provide an important basis for formulating subsequent intervention strategies. Through long-term monitoring, the system can establish the user's personalized sleep pattern, providing accurate data support for the diagnosis and intervention of sleep-wake rhythm disorders.

[0116] Through this multi-level and multi-angle signal analysis method, the system of the present invention can comprehensively and accurately capture the sleep-wake rhythm characteristics of users, laying a solid foundation for subsequent personalized intervention. This innovative analysis method not only improves the accuracy of sleep monitoring but also provides new tools and perspectives for sleep medicine research.

[0117] The system of the present invention adopts an innovative adaptive sampling technique, which not only improves the data acquisition efficiency but also ensures the accurate capture of key information. Specifically, when the body movement data is in the low amplitude segment, the system will appropriately increase the sampling interval, while when the body movement data shows a high amplitude segment, the sampling interval will be correspondingly reduced. This dynamic adjustment strategy optimizes the sampling interval to about 30% of the original sampling interval, significantly improving the data acquisition efficiency.

[0118] It is worth noting that the present invention not only applies the adaptive sampling technique to the acquisition of body movement data but also innovatively extends it to the acquisition process of PPG data. This comprehensive application greatly improves the data processing ability and response speed of the entire system. For example, when the user is in a resting state, the system may reduce the sampling rate of body movement data to 5Hz, while when the user is in an active state, it may increase to 50Hz or higher. For PPG data, the sampling rate in the resting state may be maintained at 25Hz, while in the movement state, it may increase to 100Hz to capture more subtle changes.

[0119] The intervention plan execution module 4 of the present invention is a key component of the entire system, which is responsible for selecting and executing personalized intervention strategies according to the analysis results. This module mainly includes a light therapy execution unit 41, a vibration execution unit 42, and an intervention effect evaluation unit 43.

[0120] The light therapy execution unit 41 adjusts the light intensity and duration according to the personalized intervention plan. In a preferred embodiment of the present invention, the light therapy intensity can be adjusted between 100 - 10,000 lux, and the duration is usually set within the range of 15 - 60 minutes. For example, for users with mild sleep rhythm disorders, the system may recommend receiving 2500 lux of light for 30 minutes every morning; for more severe cases, it may suggest receiving 5000 lux of light for 45 minutes in the evening. The time point of light therapy will also be dynamically adjusted according to the user's circadian rhythm to achieve the best phase shift effect.

[0121] The vibration execution unit 42 is responsible for adjusting the vibration intensity and frequency according to the personalized intervention plan. Preferably, the vibration intensity can be adjusted between 0.1 - 2.0g, and the frequency range is 10 - 200Hz. For example, for a gentle wake-up during the deep sleep stage, the system may use a vibration with an intensity of 0.2g and a frequency of 20Hz; while for a situation that requires a rapid increase in alertness, a short vibration with an intensity of 1.5g and a frequency of 150Hz may be adopted.

[0122] After the intervention effect evaluation unit 43 executes the intervention, it continuously collects the user's PPG signal and body movement data to evaluate the intervention effect. This unit adopts an innovative evaluation algorithm that combines short-term response and long-term trend analysis. The short-term response is mainly evaluated by comparing the HRV indicators and body movement patterns within 30 minutes before and after the intervention, while the long-term trend is measured by tracking the sleep efficiency and daytime function scores for 7 consecutive days. The evaluation results are fed back to the intervention plan execution module 4 for continuously optimizing the intervention strategy.

[0123] The system of the present invention further includes a data storage module 5, which is electrically connected to the signal processing module 2 and the signal analysis module 3. This module not only stores the user's original physiological data and processed feature data, but also records the user's intervention history and effect evaluation results. In an embodiment of the present invention, the data storage adopts a hierarchical structure: the original data is saved for 7 days, the processed feature data is saved for 30 days, and the intervention history and evaluation results are saved long-term. This design ensures both the integrity of the data and the optimized use of storage space.

[0124] To enhance the user experience and the usability of the system, the present invention also designs a user interface module 6. This module is electrically connected to the intervention plan execution module 4 and is mainly used to display the user's sleep-wake rhythm status and intervention suggestions, while also receiving the user's feedback and personalized settings. In a preferred embodiment of the present invention, the user interface adopts an intuitive graphical design, including visual elements such as a sleep-wake cycle graph and an intervention effect trend graph, enabling the user to easily understand their sleep condition and improvement progress.

[0125] In addition, the intervention plan execution module 4 of the present invention further includes an innovative circadian rhythm phase response curve modeling unit 44. This unit establishes an individualized circadian rhythm phase response curve (PRC) based on the user's PPG feature data. The establishment process of the PRC uses non-linear regression technology, combining multiple factors such as the light stimulation time, intensity, and the user's phase shift response. For example, the system may give the user a short light stimulation (such as 15 minutes of 5000 lux light) at different time points, and then observe the phase shift effect through continuous 7-day monitoring to draw an individualized PRC. This curve provides an important theoretical basis for formulating precise personalized intervention plans, significantly improving the pertinence and effectiveness of the intervention.

[0126] Through the above detailed description, it can be seen that the sleep-wake rhythm disorder wearable monitoring system of the present invention integrates a number of innovative technologies, from data collection, signal processing, feature extraction to personalized intervention, forming a comprehensive, efficient and intelligent closed-loop system. Such a system can not only accurately monitor and analyze the sleep-wake rhythm state of users, but also provide precise personalized intervention, which is expected to bring significant improvement in the quality of life for patients with sleep disorders.

[0127] The system of the present invention further includes a mobile terminal 7, which establishes a wireless connection with the monitoring module 1 to form a complete portable monitoring system. The mobile terminal 7 is not only used to receive and store the data collected by the monitoring module 1, but also can execute complex data processing and analysis algorithms, and display the analysis results and intervention suggestions in an intuitive manner.

[0128] In a preferred embodiment of the present invention, the mobile terminal 7 uses Bluetooth 5.0 technology to communicate with the monitoring module 1. This technology can not only ensure the stability and security of data transmission, but also significantly reduce power consumption. For example, in normal use, the battery life of the monitoring module 1 can reach more than 7 days, greatly improving the user's convenience of use.

[0129] One of the core functions of the mobile terminal 7 is to execute data processing and analysis algorithms. These algorithms include but are not limited to fast Fourier transform (FFT), wavelet transform, principal component analysis (PCA), etc. Through these algorithms, the system can extract valuable features from the original data, such as heart rate variability (HRV) indicators, sleep stage information, etc. Preferably, these computationally intensive tasks will be automatically performed when the user is resting to avoid affecting the daily use experience.

[0130] In addition, the mobile terminal 7 also undertakes the important tasks of data visualization and user interaction. The system of the present invention integrates an application based on the Android operating system on the mobile terminal 7, and this program provides rich functions and an intuitive user interface.

[0131] Specifically, this application can first visually display the sleep-wake rhythm data of the user in the form of charts, curves, etc. For example, it can generate a 24-hour activity heat map to clearly show the user's activity pattern and sleep time distribution. At the same time, the program can also draw trend charts of physiological indicators such as heart rate and body movement to help users better understand their physical conditions.

[0132] Secondly, the application provides an interface for setting personalized intervention plans. Users can adjust parameters such as the time and intensity of light therapy and vibration intervention according to their living habits and preferences. For example, users who have difficulty getting up early may choose to gradually increase the light intensity 30 minutes before the preset wake-up time, so as to achieve a more natural and comfortable wake-up process.

[0133] Furthermore, the application program of the present invention also has the function of recording and analyzing the long-term sleep-wake rhythm change trend of users. The system will automatically generate weekly and monthly reports, which detail the improvement of users' sleep quality, the implementation of intervention plans, etc. This long-term tracking not only helps users establish healthier work and rest habits but also provides valuable reference data for medical professionals.

[0134] In terms of data security, the system of the present invention has taken multiple protection measures. First of all, all transmitted data is encrypted end-to-end to ensure that information will not be intercepted by third parties. Secondly, the data stored on the mobile terminal 7 is also encrypted, so that even if the device is lost, privacy will not be leaked. In addition, users can choose to synchronize data to cloud servers, which adopt high-level security protocols such as TLS1.3 to further ensure data security.

[0135] Finally, the present invention also proposes a wearable monitoring method for sleep-wake rhythm disorder corresponding to the above system. This method includes the following steps:

[0136] First of all, the monitoring module 1 collects the body movement acceleration information, PPG signal and emotional state data of the user. In this step, the system dynamically adjusts the sampling rate according to the user's activity status to balance data accuracy and power consumption. For example, during the user's sleep, the sampling rate of the body movement sensor may be reduced to 10Hz, while during the day's activities, it may be increased to 50Hz or higher.

[0137] Next, the signal processing module 2 filters and down-samples the collected data. This step uses an innovative adaptive filtering algorithm that can effectively remove various interference signals, such as motion artifacts, environmental light interference, etc. The quality of the processed data directly affects the accuracy of subsequent analysis, so this step plays a key role in the whole method.

[0138] Then, the signal analysis module 3 uses the pulse neural network described in detail above to analyze the processed body movement acceleration information and PPG signal to obtain the phase and circadian rhythm deviation of the individual PPG signal. This step can not only identify the user's sleep-wake cycle but also detect subtle rhythm disorders, providing an accurate basis for subsequent intervention.

[0139] After obtaining these key pieces of information, the system will optimize the individual's intervention strategy using a reinforcement learning algorithm based on Q-learning. This process first inputs the individual's classification features into the intervention module to select personalized strategies for light therapy or vibration. Then, it combines the time series and event analysis modules to evaluate nighttime awakenings, while using the reinforcement learning algorithm to analyze the individual's sleep efficiency and the number of awakenings. Finally, the system calculates the accuracy of phase deviation correction and uses it as the input to the reward function to continuously optimize the intervention strategy.

[0140] Based on the optimized strategy, the system implements a personalized intervention plan, including light therapy and vibration interventions. The intensity, timing, and duration of these interventions are precisely adjusted according to the individual's situation to achieve the best results.

[0141] Finally, the system continuously evaluates the intervention effect and updates the intervention strategy according to the evaluation results. This closed-loop feedback mechanism ensures that the intervention plan can be continuously optimized as the user's condition changes, always maintaining the best effect.

[0142] From the above detailed description, it can be seen that the wearable monitoring system and method for sleep-wake rhythm disorder provided by the present invention have high innovation and practicality. It can not only accurately monitor and analyze the sleep-wake rhythm state of users, but also provide personalized intervention plans and continuously optimize these plans through intelligent algorithms. This comprehensive solution is expected to bring significant improvements in the quality of life for sleep disorder patients and provide valuable data support for research in related fields.

[0143] To verify the effectiveness and superiority of the wearable monitoring system and method for sleep-wake rhythm disorder of the present invention, a controlled experiment lasting 8 weeks was designed. The experiment recruited 60 volunteers diagnosed with mild to moderate sleep-wake rhythm disorder, aged 25 - 55 years old, with an equal number of men and women. The experiment adopted a randomized controlled double-blind design, randomly dividing the volunteers into three groups of 20 people each.

[0144] Experimental group: Use the complete system of the present invention, including an intelligent wristband, a mobile terminal application, and a personalized intervention plan.

[0145] Control group 1: Use common sleep monitoring bracelets on the market and the corresponding mobile phone applications, but do not provide personalized intervention plans.

[0146] Control group 2: Only receive regular sleep hygiene education and advice, without using any wearable devices.

[0147] Before the experiment began, all participants underwent a comprehensive sleep assessment, including the Pittsburgh Sleep Quality Index (PSQI) questionnaire, sleep log recording, and a one-week monitoring with a wrist actigraph. During the experiment, participants were evaluated every two weeks, and a final evaluation was conducted at the end of the experiment.

[0148] The main evaluation indicators included:

[0149] 1. Sleep quality: Scored using the PSQI, with a score range of 0 - 21. The lower the score, the better the sleep quality.

[0150] 2. Sleep efficiency: Calculated by the wrist actigraph, the percentage of sleep time in bed time.

[0151] 3. Daytime function: Evaluated using the Epworth Sleepiness Scale (ESS), with a score range of 0 - 24. The lower the score, the better the daytime wakefulness.

[0152] 4. Sleep - wake cycle stability: Using the relative amplitude (RA) index in non - parametric circular analysis (NPCRA). The closer the RA value is to 1, the more stable the circadian rhythm.

[0153] 5. User satisfaction: Evaluated using a custom 5 - point Likert scale. The higher the score, the higher the satisfaction.

[0154] The experimental results are shown in the following table:

[0155] Index Example group Control group 1 Control group 2 Improvement in PSQI score -5.8±1.2 -3.2±0.9 -1.5±0.7 Improvement in sleep efficiency 12.5％±2.1％ 7.8％±1.8％ 3.2％±1.5％ Improvement in ESS score -6.3±1.1 -3.7±0.8 -1.9±0.6 Improvement in RA value 0.18±0.03 0.09±0.02 0.04±0.02 User satisfaction 4.6±0.3 3.8±0.4 2.9±0.5

[0156] The experimental results show that the system of the present invention is significantly superior to the other two groups in terms of improving sleep quality, increasing sleep efficiency, improving daytime function, stabilizing the sleep - wake cycle, and user satisfaction. Notably, in the improvement of sleep - wake cycle stability (RA value), the system of the present invention shows a significant advantage, which directly verifies the effectiveness of the present invention in regulating the sleep - wake rhythm.

[0157] Further analysis found that the improvement effect of the example group began to be significantly better than the other two groups in the 4th week of the experiment, and this advantage became more obvious in the later stage of the experiment. This indicates that the system of the present invention can not only produce effects quickly but also continuously optimize the intervention plan to achieve long - term and stable improvement.

[0158] In particular, an in-depth analysis was conducted on the top 5 participants in the experimental group, and it was found that their common feature was strict adherence to the intervention plan recommended by the system and active adjustment of daily living habits. The PSQI scores of these participants decreased by an average of 7.2 points, the sleep efficiency increased by 15.8%, the ESS scores decreased by 7.5 points, and the RA value increased by 0.22. This set of data can be regarded as the best embodiment of the system of the present invention under ideal usage conditions.

[0159] These results fully demonstrate the effectiveness of several key innovations of the system of the present invention:

[0160] 1. The superiority of the adaptive hybrid pulse neural network in capturing the characteristics of the sleep-wake rhythm is reflected in the significant improvement of the RA value.

[0161] 2. The effect of the reinforcement learning algorithm based on Q-learning in optimizing personalized intervention strategies is reflected in the continuous improvement trend of various indicators.

[0162] 3. The adaptive sampling technology and real-time feedback mechanism improve the user experience, which is verified in the user satisfaction score.

[0163] 4. The comprehensive effect of the comprehensive monitoring and intervention plan on improving sleep quality and daytime function is reflected in the significant improvement of the PSQI and ESS scores.

[0164] Generally speaking, this experiment strongly proves the superiority and innovation of the present invention in improving sleep-wake rhythm disorders. It not only performs excellently in objective indicators but also receives high recognition from users, showing good application prospects. This comprehensive solution combining advanced algorithms, precise monitoring, and personalized intervention provides a new research direction and practical tool for the field of sleep medicine.

[0165] It should be noted that the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A wearable monitoring system for sleep-wake rhythm disorders, characterized in that: include: Monitoring module for: Collect the user's body acceleration information and PPG signal; Detecting the user's emotional state data; The signal processing module is electrically connected to the monitoring module and is used to: Receiving body acceleration information, PPG signal and emotional state data sent by the monitoring module; Filtering and down-sampling the body motion acceleration information, PPG signal and emotional state data; A signal analysis module is electrically connected to the signal processing module and is used to: Receiving the processed body acceleration information and PPG signal sent by the signal processing module; Analyzing the processed body acceleration information and PPG signal using a spiking neural network to obtain time domain and non-time domain features; The intervention plan execution module is electrically connected to the signal analysis module and is used to: Receiving the time domain and non-time domain features sent by the signal analysis module; Selecting a personalized intervention plan using a Q-learning-based reinforcement learning algorithm based on the phase deviation, nighttime sleep arousal, and sleep efficiency data in the temporal and non-temporal features; The personalized intervention regimen is performed, wherein the personalized intervention regimen includes light therapy and vibration.

2. The wearable monitoring system for sleep-wake rhythm disorders according to claim 1, characterized in that: The signal analysis module comprises: PPG signal time series feature extraction unit is used to: Calculate the phase deviation of the PPG signal; Providing prior knowledge for the assessment of individual nighttime sleep efficiency; According to the nighttime sleep assessment results, the accuracy of phase correction, sleep efficiency, and number of awakenings were obtained as reward functions.

3. The wearable monitoring system for sleep-wake rhythm disorders according to claim 1, characterized in that: The spiking neural network is an adaptive hybrid spiking neural network, comprising: Time-domain based hybrid pulse model for: Use pulse counting method to count the time characteristics of time series; Provides methods for quantifying differences in pulse occurrence times; Scaled Normalized Adaptive Pulse Model for: Overcoming the linear relationship of time series; Dynamically adjust the activation threshold of neurons according to the training process; Adjust the pulse width based on the standard deviation and sample entropy of the input time series; A periodic pulse model based on the time-frequency domain is used for: The pulse time interval of the time series is measured by Fourier transform and time correlation entropy; A scale-normalized periodic pulse model is adopted.

4. The wearable monitoring system for sleep-wake rhythm disorders according to claim 1, characterized in that: The system uses adaptive sampling technology to: Increase the sampling interval when the body motion data is in the low amplitude segment; When the body motion data is in the high amplitude segment, reduce the sampling interval; Optimize the sampling interval to 30% of the original sampling interval; The adaptive sampling technology is applied to PPG data collection.

5. The wearable monitoring system for sleep-wake rhythm disorder according to claim 1, characterized in that: The intervention program execution module includes: Light therapy delivery unit for: adjusting light intensity and duration according to the personalized intervention plan; performing light therapy interventions over a preset time period; Vibration actuator for: adjusting the vibration intensity and frequency according to the personalized intervention plan; Perform vibration intervention within a preset time; Intervention effectiveness evaluation unit, used to: Collect PPG signals and body motion data after intervention; Evaluate the intervention effect and provide feedback to the intervention program execution module for strategy optimization.

6. The wearable monitoring system for sleep-wake rhythm disorder according to claim 1, characterized in that: Also includes: The data storage module is electrically connected to the signal processing module and the signal analysis module, and is used to: Store the user's original physiological data and processed feature data; Record the user's intervention history and effect evaluation results; A user interface module is electrically connected to the intervention program execution module and is used to: Display the user's sleep-wake rhythm status and intervention suggestions; Receive user feedback and personalize settings.

7. The wearable monitoring system for sleep-wake rhythm disorders according to claim 1, characterized in that: The intervention program execution module also includes: Circadian phase response curve modeling unit for: Establish a personalized circadian rhythm phase response curve based on the user's PPG characteristic data; Provide a theoretical basis for the formulation of personalized intervention plans.

8. The wearable monitoring system for sleep-wake rhythm disorders according to claim 1, characterized in that: The system further comprises: The mobile terminal is wirelessly connected to the monitoring module and is used to: Receiving and storing data collected by the monitoring module; implement data processing and analysis algorithms; Displays analysis results and intervention recommendations.

9. The wearable monitoring system for sleep-wake rhythm disorders according to claim 8, characterized in that: The mobile terminal includes an application based on the Android operating system, which is used to: Visualize the user's sleep-wake rhythm data; Provide a personalized intervention plan setting interface; Record and analyze the user's long-term sleep-wake rhythm change trends.

10. A wearable monitoring method for sleep-wake rhythm disorders, based on the system according to any one of claims 1 to 9, characterized in that: The following steps are involved: S1: Collect the user's body acceleration information, PPG signal and emotional state data through the monitoring module; S2: The signal processing module filters and downsamples the collected data, and the filtered data is input into the signal analysis module; S3: The signal analysis module uses the pulse neural network to analyze the processed body acceleration information and PPG signal to obtain the phase and circadian rhythm deviation of the individual PPG signal; S4: Use the Q-learning-based reinforcement learning algorithm to optimize the individual intervention strategy. The optimization of the intervention strategy is achieved through the following steps: The individual classification characteristics were input into the intervention module to select the personalized strategy of light therapy or vibration; Combined time series and event analysis modules to assess nighttime awakenings; Combined with reinforcement learning algorithms to analyze individual sleep efficiency and wake-up times; Calculate the accuracy of phase deviation correction as input of reward function; S5: Implement personalized intervention plans, including light therapy and vibration interventions; S6: Evaluate intervention effectiveness and update intervention strategies.

Images