Multi-mode collaborative bracelet awakening system and method based on sleep stage perception

Abstract

According to the sleep stage perception-based multi-mode collaborative bracelet awakening system provided by the invention, a user is mildly awakened in combination with a multi-mode progressive mode by judging the transfer path of the sleep stage. According to the system, the hidden Markov model is improved, the multi-modal physiological features are fused, and the stage recognition precision is improved. The problem that a user is poor in spirit due to the fact that a traditional intelligent bracelet is forcibly awakened at fixed time is solved, and the mental state and comfort of the awakened user are effectively improved.

Description

Technical Field

[0001] This application relates to the field of smart wearable technology, specifically to a multimodal collaborative wristband wake-up system and method based on sleep stage perception. Background Technology

[0002] Common sense suggests that the longer you sleep, the more refreshed you will be when you wake up. While this applies to specific groups, it's not entirely accurate when generalized to the general population. Modern young people often experience longer sleep onset times and shorter sleep durations due to work pressure and late nights, so appropriately increasing sleep duration can help them achieve a more refreshed state, thus validating the aforementioned common sense. However, for naturally long sleepers or those with regular sleep patterns, simply increasing sleep duration is not conducive to a good mental state and can easily lead to the common experience of "feeling more tired the more you sleep."

[0003] A person's mental state is closely related to the integrity of their sleep cycle and the timing of awakening. For example, it's easier to fall back asleep if you wake up in the first half of the night than in the second half. When people wake up in the first half of the night, they feel like they haven't slept enough and then quickly fall back asleep. The professional explanation for "not having slept enough" is that in the first half of the night, after falling asleep, deep sleep accounts for a higher proportion. When a person wakes up during a deep sleep cycle, they feel tired. However, when a person wakes up in the second half of the night, they feel very awake and often stay awake until dawn or for an hour or two before they can fall back asleep. This is because when a person wakes up during a light sleep stage, they are more likely to enter a state of wakefulness.

[0004] Based on the above, the integrity of the sleep cycle and the timing of wakefulness jointly affect mental state upon waking. Therefore, a full mental state requires sufficient sleep cycles and appropriate wakefulness timing. Smart bracelets, as a recent hot topic, focus primarily on monitoring sleep quality. However, this monitoring lacks guidance and may even create a burden for falling asleep the next day, as people experience anxiety when they know their sleep quality is poor but cannot change it. Furthermore, the wake-up function of current smart bracelets largely replaces the traditional alarm clock, lacking true "intelligence." In addition, vibration and sound are sudden, strong stimuli for waking, far less gentle than light. Summary of the Invention

[0005] This application provides a multimodal collaborative wristband wake-up system based on sleep stage perception, which solves the problem of inappropriate wake-up timing caused by the forced wake-up at a fixed time in traditional smart wristbands.

[0006] The technical solution of this application is as follows: A multimodal collaborative wristband wake-up system based on sleep stage perception allows the user to set a wake-up time interval. The system includes a vital sign monitoring module, a data processing module, and a wake-up module. The vital sign monitoring module acquires the user's sleep vital signs and, based on these signs, outputs the user's sleep stage through the data processing module. The user's sleep signs are characterized by a three-dimensional observation vector consisting of the LF / HF ratio, body motion entropy, and blood oxygen fluctuation slope. The data processing module periodically records the user's sleep stages and normalizes consecutive identical sleep stages to obtain a sleep stage sequence. The data processing module traverses the sleep stage sequence and queries the standard sequence of the sleep stage with the highest frequency as a comparison sequence. The standard sequence of sleep stages is defined as a fixed number of consecutive and different sleep stages combined. The data processing module determines whether the comparison sequence contains the N1 stage based on the position of the sleep stage at the start of the wake-up time interval in the comparison sequence. If so, the user's sleep stage will be continuously monitored and a command will be sent by the wake-up module to wake the user in the next N1 stage; If not, the user's sleep stage is continuously monitored, and when the N1 stage occurs, the wake-up module sends a command to wake the user.

[0007] Furthermore, when determining whether the comparison sequence contains the N1 stage, if the comparison sequence does not include the sleep stage at the beginning of the wake-up time interval, then it is determined whether the comparison sequence contains the N1 stage. If so, the user's sleep stage will be continuously monitored and a command will be sent by the wake-up module to wake the user when the second N1 stage occurs; If not, the user's sleep stage is continuously monitored, and when the N1 stage occurs, the wake-up module sends a command to wake the user.

[0008] Furthermore, in the backward comparison sequence to determine whether the negative option of the N1 stage is included, if the N1 stage does not appear in the wake-up time interval, the wake-up module sends an instruction to wake up the user at the end of the wake-up time interval.

[0009] Secondly, this application also provides a multimodal collaborative wristband wake-up system based on sleep stage perception. The user sets a wake-up time interval. The wristband wake-up system includes a vital sign monitoring module, a data processing module, and a wake-up module. The vital sign monitoring module acquires the user's sleep vital signs and, based on these signs, outputs the user's sleep stage through the data processing module. The user's sleep signs are characterized by a three-dimensional observation vector consisting of the LF / HF ratio, body motion entropy, and blood oxygen fluctuation slope. The data processing module periodically records the user's sleep stages and normalizes consecutive identical sleep stages to obtain a sleep stage sequence. The data processing module traverses the sleep stage sequence and queries the standard sequence of the sleep stage with the highest frequency as a comparison sequence. The standard sequence of sleep stages is defined as a fixed number of consecutive and different sleep stages combined. The data processing module determines whether the comparison sequence contains the N1 stage based on the position of the sleep stage at the start of the wake-up time interval in the comparison sequence, both at the current position and the next position. If so, the user's sleep stage will be continuously monitored and the wake-up module will send a command to wake the user when the first N1 stage occurs; If not, determine whether the current and subsequent sleep stages include stage N2. If yes, continue monitoring the user's sleep stages and send a command to wake the user when stage N2 occurs. If not, continue monitoring the user's sleep stages and send a command to wake the user when stage N1 or stage N2 is detected first. If neither stage N1 nor stage N2 is detected, send a command to wake the user when the wake-up time interval is reached.

[0010] Furthermore, when determining whether the comparison sequence contains the N1 stage, both currently and subsequently, if the comparison sequence does not include the sleep stage at the start of the wake-up time interval, then it is determined whether the comparison sequence contains the N1 stage. If so, the user's sleep stage will be continuously monitored and the wake-up module will send a command to wake the user when the next N1 stage occurs; If not, the user's sleep stage is continuously monitored, and when the N1 stage occurs, the wake-up module sends a command to wake the user.

[0011] Furthermore, when determining whether the comparison sequence contains the N1 stage, both currently and subsequently, if the comparison sequence does not include the sleep stage at the start of the wake-up time interval, then it is determined whether the comparison sequence contains the N1 stage. If the N1 phase does not occur during the wake-up interval, the wake-up module sends a command to wake up the user at the end of the wake-up interval.

[0012] Furthermore, the wake-up time interval is greater than or equal to 30 minutes, and the vital signs monitoring module monitors the sleep stage every 30 seconds.

[0013] Furthermore, the wake-up module includes a tactile vibration module, a lighting control module, and a bone conduction speaker module; The tactile vibration module includes a linear enhancement segment, a pulse vibration segment, and a random vibration segment; The lighting control module is wirelessly connected to the lighting fixture and controls the lighting fixture to increase from the initial temperature measurement to the preset color temperature during the pulse vibration segment and the random vibration segment.

[0014] Furthermore, the color temperature variation function of the lamp is as follows: In the formula, T ( t () indicates time t The color temperature value; Indicates the initial color temperature; Indicates the target color temperature; k Indicates the slope of the growth rate; This indicates the time at which the color temperature changes.

[0015] Furthermore, the bone conduction speaker module increases from the initial decibel level to the preset decibel level after the random vibration segment ends; The volume change function of a bone conduction loudspeaker is: ; In the formula, Indicates the initial volume; Indicates the target volume; Indicates the duration of growth.

[0016] Due to the adoption of the above technical solution, the beneficial effects of this application are as follows: 1. This application provides two technical solutions for the majority of users of the smart bracelet, each corresponding to a different wake-up habit. This application predicts sleep transition paths based on a comparison sequence derived from changes in nighttime sleep stages. The sleep transition path represents the change in sleep stages. Normal sleep stages exhibit a clear periodic change; this application leverages this periodicity and an accurate sleep stage prediction model to predict sleep stage transition paths. The transition path does not have temporal characteristics; it only indicates the order of transitions between sleep stages and the number of sleep stages included in the path. Therefore, the comparison sequence is directional; the current position represents the sleep stage at the start of wake-up, and subsequent positions represent subsequent stages of the current sleep stage. The comparison sequence can be set to include at least three sleep stages to cover most cases of sequential sleep stage changes.

[0017] The first mode is the snooze mode, suitable for users who are not sensitive to wake-up time, i.e., when the wake-up time interval is relatively long. In this mode, a sleep stage determination is performed when the wake-up time interval is reached, and then the user is woken up when the second N1 stage occurs, in order to extend the user's sleep time while ensuring that they wake up at the appropriate time, thus ensuring that the user has sufficient energy. The vital signs monitoring module periodically determines the sleep stage. During this process, the data processing module integrates consecutive identical sleep stages. Therefore, the second N1 stage does not refer to the second sleep stage monitored by the vital signs monitoring module. The second N1 stage refers to the N1 stage after one sleep stage transition.

[0018] The second mode is the normal mode, suitable for users who wake up on time. In this mode, the goal is to ensure the user wakes up at the appropriate time. The wake-up process begins when the user's sleep stage is detected as N1. If the transition path does not include stage N1, the system aims to wake the user during stage N2. This mode is suitable for users who wake up on time, and also helps users who stay up late to wake up healthily, as their deep sleep (N3) stage is delayed, and waking up during N3 is avoided as much as possible.

[0019] 2. This application uses both sound and light to gently wake the user, minimizing the risk of startling them and helping them regain more energy. Furthermore, bone conduction is gentler than traditional external speakers and will not disturb others.

[0020] 3. This application improves upon traditional sleep stage prediction. Traditional Hidden Markov Models (HMMs) typically construct observation sequences using single-channel or low-dimensional physiological signals (such as heart rate or body movement), resulting in simple observation probability modeling but failing to characterize the physiological differences between complex states. This application improves upon traditional HMMs by replacing the single Gaussian modeling method with a Gaussian Mixture Model (GMM) using three-dimensional vital sign parameters, thereby enhancing the modeling of observation probabilities and improving the accuracy of sleep stage prediction. Attached Figure Description

[0021] The accompanying drawings, which are provided to further illustrate this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application.

[0022] Figure 1 A flowchart of a multimodal collaborative wristband wake-up system based on sleep stage perception is provided for this application; Figure 2 A flowchart of another multimodal collaborative wristband wake-up system based on sleep stage perception provided in this application. Detailed Implementation

[0023] It should be noted that this application is based on the regular sleep cycle transition. Numerous studies have shown that the sleep stages of normal individuals change cyclically; therefore, the sleep stage transition path in this application reflects the changes in the user's sleep state.

[0024] Example: Data Acquisition and Preprocessing: Raw physiological data, including heart rate, body movement, and blood oxygen saturation, are collected every 30 seconds and preprocessed in real time. Based on the acquired physiological data, feature extraction algorithms are used to calculate frequency domain features, body movement data entropy, and SpO2 change rate. Specifically, HRV frequency domain feature extraction uses Fast Fourier Transform (FFT) to perform frequency domain analysis on the RR interval sequence, obtaining low-frequency (LF) and high-frequency (HF) components, and calculating the LF / HF ratio as an indicator of autonomic nervous system activity. In the formula, This indicates the power in the 0.04–0.15Hz range; This represents the power in the 0.15–0.40 Hz range. The body entropy is calculated using a 30-second acceleration window, with the sample entropy (SampEn) used as an indicator of the complexity of the body motion amplitude. ; In the formula, m The embedding dimension represents the number of consecutive sampling points in each comparison (2 in this embodiment). r The similarity tolerance threshold is defined as 0.2 times the standard deviation of the sequence. N The length of the acceleration sequence depends on the accelerometer sampling rate and time window (e.g., 30 seconds × 50 Hz sampling rate, N = 1500); A and B represent the number of m+1 dimensional and m dimensional vector pairs that satisfy the condition under the similarity tolerance threshold r, respectively.

[0025] The slope of blood oxygen fluctuation is calculated by taking the first-order linear fit slope of the SpO2 sequence of the previous 30 seconds at the current moment. K 1) Obtain: ; In the formula, 1 indicates the linear trend (i.e., the slope of change) of SpO2 fluctuations. Indicates the first i The timestamps corresponding to each sampling point; Indicates the first i The blood oxygen value corresponding to each sampling point; This represents the average value over the sampling time. Mean blood oxygen level.

[0026] This application collects three types of physiological characteristics of the user every 30 seconds, including the LF / HF ratio, body motion sample entropy (SampEn), and SpO2 change rate, which constitute a three-dimensional feature observation vector: The feature observation vectors are input into a Gaussian Mixture Model (GMM)-Hidden Markov Model (HMM) to determine the current sleep stage. The model sets five hidden states. : Wake, N1, N2, N3, REM. Each hidden state The following observation probability distributions are all modeled using a multimodal Gaussian mixture model, namely: Where K is the number of Gaussian components. The weight of the k-th component (satisfying) ), The mean of the components (three-dimensional vector). Let covariance matrix be the variance matrix. It is a multivariate normal density function. This modeling method can fully describe the joint distribution characteristics of multimodal physiological signals in each sleep stage, achieving more accurate state discrimination.

[0027] The above model is used to characterize the joint distribution of multimodal physiological signals in the current state.

[0028] The specific implementation path is as follows: First, for each hidden state... The next k Each Gaussian component is initialized with parameters, including weights. ,satisfy Mean Covariance Then calculate the degree of responsibility for the state. Each training sample below Calculate the probability (responsibility) of its generation by the k-th Gaussian component. : ; The above formula gives the state to which the t-th observation belongs. In this case, the "probability ratio" generated by the k-th Gaussian component. The more "concentrated" or "narrow" a Gaussian component is, the more regular the characteristic changes in that sleep state are; the more "flat" or "biased" it is, the greater the fluctuation or bias in the physiological data of that state.

[0029] Next, based on all observation points at the 1st Based on the level of responsibility under the current state, update the parameters as follows: Calculate the first k Number of valid samples for each Gaussian component: ; Update the mixed weights: ; Update the mean vector: ; Update the covariance matrix: ; The above updates are all weighted averages and weighted covariances, with the weights derived from the responsibility level calculated above. .

[0030] Finally, convergence is checked, and the log-likelihood value of the current round is calculated: If the increase in the log-likelihood value is less than the set threshold, then convergence is considered; otherwise, proceed to the next round n+1.

[0031] The Gaussian density function is as follows: ; In the formula, μ It is the mean vector. Σ Let covariance matrix be the variance matrix. d The feature dimension is defined as follows. The output value of this function represents the probability density of each observation point (a set of three-dimensional physiological feature values ​​collected and extracted every 30 seconds, representing the user's physiological state at that moment) belonging to a certain Gaussian component, which is used to calculate the degree of responsibility and parameter updates.

[0032] The system extracts three features—heart rate variability (LF / HF), body kinetic entropy (SampEn), and blood oxygen saturation slope (SpO2Slope)—within a 30-second time window, and identifies the user's current sleep stage using a combined threshold rule. To improve the reproducibility of the technology, the discrimination criteria are defined using a numerical range as follows: Wake period: LF / HF > 2.0, SampEn > 1.5, SpO2 slope > 0.001; Light sleep (N1): LF / HF∈[1.5,2.0], SampEn∈[1.0,1.5], SpO2 slope∈[-0.001,0.001]; Mid-deep sleep (N2 / N3): LF / HF∈[1.0,1.5], SampEn<1.0, SpO2 slope∈[-0.0005,0.0005]; Rapid eye movement (REM): LF / HF>2.0, SampEn>1.2, mean SpO2 ≥95%, slope ∈[0,0.0015].

[0033] The data processing module records sleep stages and merges consecutive identical sleep stages into one stage for normalization. Upon reaching the wake-up time interval, a sleep stage sequence S is obtained. S contains all sleep stages for the night, and adjacent sleep stages are distinct. The comparison sequence is defined as a tuple, meaning each comparison sequence includes 3 sleep stages. The module iterates through combinations of 3 consecutive sleep stages in S, selecting the one with the highest frequency as the final comparison sequence. For example, if S contains 10 sleep stages (S1-S2-S3), (S2-S3-S4)...(S8-S9-S10), comparison sequences (S1-S2-S3), (S2-S3-S4),...(S8-S9-S10) will be formed. The sequence with the highest frequency is selected as the comparison sequence. Upon reaching the wake-up interval, the system calls the currently output... At the current stage Check if it is in If not The middle part is directly judged. Does it include stage N1? Because the comparison sequence reflects the individual's sleep habits that night, if the current sleep stage is not included... In the middle, it can be based on The system checks for the presence of N1 sleep stages to trigger wake-up. If N1 is absent, wake-up is initiated based on monitoring results. If N1 is present, it indicates that N1 may occur again in the future, requiring continued monitoring and wake-up. In most cases, the number of sleep stage transitions each night is not too frequent, and the N3 (deep sleep) stage usually does not occur in the morning. Therefore, the comparison sequence of 3-tuples is sufficient to cover the commonly occurring sleep stages—N2, N1, and REM.

[0034] if exist In this process, wake-up is performed according to two modes. The two technical solutions correspond to two different wake-up habits. This application predicts sleep transition paths based on a comparison sequence derived from changes in nighttime sleep stages. A sleep transition path is essentially a change in sleep stages. Normal sleep stages exhibit significant periodic changes. This application utilizes the periodicity of sleep and an accurate sleep stage prediction model to predict sleep stage transition paths. The transition path does not have temporal characteristics; it only indicates the order of transitions between sleep stages and the number of sleep stages included in the transition path. Therefore, the comparison sequence is directional; the current position represents the sleep stage at the start of wake-up, and subsequent positions represent subsequent stages of the current sleep stage. The comparison sequence can be set to include at least three sleep stages to cover most cases of sequential sleep stage changes. (See attached...) Figure 1 and attached Figure 2 As shown.

[0035] 1. Snooze mode The first mode is the snooze mode, suitable for users who are not sensitive to wake-up time, i.e., when the wake-up time interval is relatively long. In this mode, a sleep stage determination is performed when the wake-up time interval is reached, and then the user is woken up when the second N1 stage appears, in order to extend the user's sleep time while ensuring that they wake up at the appropriate time, thus ensuring that the user has sufficient energy. When checking whether the comparison sequence contains the N1 stage, if the comparison sequence does not include the sleep stage at the beginning of the wake-up time interval, then it is determined whether the comparison sequence contains the N1 stage. If so, the user's sleep stage will be continuously monitored and a command will be sent by the wake-up module to wake the user when the second N1 stage occurs; If not, the system continuously monitors the user's sleep stages and sends a command to wake the user when stage N1 occurs. The vital signs monitoring module periodically determines the sleep stages. During this process, the data processing module integrates consecutive identical sleep stages. Therefore, the second N1 stage does not refer to the second sleep stage monitored by the vital signs monitoring module. The second N1 stage refers to the N1 stage after one sleep stage transition.

[0036] In the negative option of determining whether the comparison sequence contains the N1 stage, if the N1 stage does not appear in the wake-up time interval, the wake-up module sends a command to wake up the user at the end of the wake-up time interval.

[0037] 2. Normal Mode This is suitable for users who are sensitive to wake-up time, such as those using "fixed wake-up time" or "weekday mode." The normal mode also determines the sleep stage at the start and outputs a comparison sequence. If the current sleep stage is N1, a wake-up is performed. If the current sleep stage is not N1, the system checks if N1 exists in the subsequent transition path. If N1 does not exist, it checks if N2 exists. If N2 exists, the system monitors the sleep stage and wakes the user at N2. If N2 is not included in the predicted sleep transition path, the system continues to monitor the sleep stage, abandons predicting the sleep transition path, and wakes the user at N1 in order of priority, followed by N2, and finally the latest wake-up time. In practice, it is impossible to determine whether N1 will exist after N2. In this case, N1 and N2 have the same priority: the first detected N1 stage is woken up, the first detected N2 stage is also woken up, and if neither is detected, the latest wake-up time is used. The N2 stage is considered the second priority wake-up point because its physiological indicators are between deep sleep and light sleep, and the wake-up experience is relatively acceptable. If the current stage is REM, the system defaults to waiting for the stage transition and does not wake the user immediately.

[0038] When determining whether the comparison sequence contains the N1 stage, both currently and subsequently, if the comparison sequence does not include the sleep stage at the start of the wake-up time interval, then determine whether the comparison sequence contains the N1 stage. If so, the user's sleep stage will be continuously monitored and the wake-up module will send a command to wake the user when the next N1 stage occurs; If not, the user's sleep stage is continuously monitored, and when the N1 stage occurs, the wake-up module sends a command to wake the user.

[0039] In normal mode, if the N1 phase does not occur in the wake-up interval, the wake-up module sends a command to wake up the user at the end of the wake-up interval.

[0040] In practice, the smart wristband platform uses an ESP32 main control chip to control the embedded vibration motor, bone conduction speaker, and lighting control module.

[0041] This embodiment uses a wake-up module to perform the wake-up process. The wake-up module includes a tactile vibration module, a lighting control module, and a bone conduction speaker module. These three modules do not operate simultaneously. First, a gentle wake-up is performed by the random vibration module, as people are less sensitive to touch during sleep. Then, the lighting control module wirelessly connects to the smart lighting fixtures via Bluetooth or Wi-Fi to control external smart lighting devices in the user's bedroom (such as smart bedside lamps or table lamps), enabling remote adjustment of color temperature and brightness. The module adjusts the brightness of the lighting fixtures during the process.

[0042] The color temperature change function of the lamp is as follows: In the formula, T ( t () indicates time t The color temperature value; Indicates the initial color temperature; Indicates the target color temperature; k Indicates the slope of the growth rate; This indicates the time at which the color temperature changes.

[0043] The bone conduction speaker module increases from the initial decibel level to the preset decibel level after the random vibration segment ends; The volume change function of a bone conduction loudspeaker is: ; In the formula, Indicates the initial volume; Indicates the target volume; Indicates the duration of growth.

[0044] In practice, the tactile vibration module starts working first, lasting for 30 seconds. If the user turns off the wristband notification, the bone conduction speaker module is activated. While these two modules are operating, the lighting control module gradually increases the color temperature of the indoor lights.

[0045] For any parts not mentioned in this application, existing technologies may be used or referenced.

[0046] The above description is merely an embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of this application should be included within the scope of the claims of this application.

Claims

1. A multimodal collaborative wristband wake-up system based on sleep stage perception, wherein the user sets a wake-up time interval, the wristband wake-up system includes a vital sign monitoring module, a data processing module, and a wake-up module, wherein the vital sign monitoring module acquires the user's sleep vital signs and outputs the user's sleep stage based on the sleep vital signs through the data processing module, characterized in that, The user's sleep signs are characterized by a three-dimensional observation vector consisting of the LF / HF ratio, body motion entropy, and blood oxygen fluctuation slope. The data processing module periodically records the user's sleep stages and normalizes consecutive identical sleep stages to obtain a sleep stage sequence. The data processing module traverses the sleep stage sequence and queries the standard sequence of the sleep stage with the highest frequency as a comparison sequence. The standard sequence of sleep stages is defined as a fixed number of consecutive and different sleep stages combined. The data processing module determines whether the comparison sequence contains the N1 stage based on the position of the sleep stage at the start of the wake-up time interval in the comparison sequence. If so, the user's sleep stage will be continuously monitored and a command will be sent by the wake-up module to wake the user in the next N1 stage; If not, the user's sleep stage is continuously monitored, and when the N1 stage occurs, the wake-up module sends a command to wake the user.

2. The multimodal collaborative wristband wake-up system based on sleep stage perception according to claim 1, characterized in that, When determining whether the comparison sequence contains the N1 stage, if the comparison sequence does not include the sleep stage at the beginning of the wake-up time interval, then determine whether the comparison sequence contains the N1 stage. If so, the user's sleep stage will be continuously monitored and a command will be sent by the wake-up module to wake the user when the second N1 stage occurs; If not, the user's sleep stage is continuously monitored, and when the N1 stage occurs, the wake-up module sends a command to wake the user.

3. The multimodal collaborative wristband wake-up system based on sleep stage perception according to claim 2, characterized in that, In the negative option of determining whether the comparison sequence contains the N1 stage, if the N1 stage does not appear in the wake-up time interval, the wake-up module sends a command to wake up the user at the end of the wake-up time interval.

4. A multimodal collaborative wristband wake-up system based on sleep stage perception, wherein the user sets a wake-up time interval, the wristband wake-up system includes a vital sign monitoring module, a data processing module, and a wake-up module, wherein the vital sign monitoring module acquires the user's sleep vital signs and outputs the user's sleep stage based on the sleep vital signs through the data processing module; characterized in that, The user's sleep signs are characterized by a three-dimensional observation vector consisting of the LF / HF ratio, body motion entropy, and blood oxygen fluctuation slope. The data processing module periodically records the user's sleep stages and normalizes consecutive identical sleep stages to obtain a sleep stage sequence. The data processing module traverses the sleep stage sequence and queries the standard sequence of the sleep stage with the highest frequency as a comparison sequence. The standard sequence of sleep stages is defined as a fixed number of consecutive and different sleep stages combined. The data processing module determines whether the comparison sequence contains the N1 stage based on the position of the sleep stage at the start of the wake-up time interval in the comparison sequence, both at the current position and the next position. If so, the user's sleep stage will be continuously monitored and the wake-up module will send a command to wake the user when the first N1 stage occurs; If not, determine whether the current and subsequent sleep stages include stage N2. If yes, continue monitoring the user's sleep stages and send a command to wake the user when stage N2 occurs. If not, continue monitoring the user's sleep stages and send a command to wake the user when stage N1 or stage N2 is detected first. If neither stage N1 nor stage N2 is detected, send a command to wake the user when the wake-up time interval is reached.

5. A multimodal collaborative wristband wake-up system based on sleep stage perception according to claim 4, characterized in that, When determining whether the comparison sequence contains the N1 stage, both currently and subsequently, if the comparison sequence does not include the sleep stage at the start of the wake-up time interval, then determine whether the comparison sequence contains the N1 stage. If so, the user's sleep stage will be continuously monitored and the wake-up module will send a command to wake the user when the next N1 stage occurs; If not, the user's sleep stage is continuously monitored, and when the N1 stage occurs, the wake-up module sends a command to wake the user.

6. A multimodal collaborative wristband wake-up system based on sleep stage perception according to claim 5, characterized in that, When determining whether the comparison sequence contains the N1 stage, both currently and subsequently, if the comparison sequence does not include the sleep stage at the start of the wake-up time interval, then determine whether the comparison sequence contains the N1 stage. If the N1 phase does not occur during the wake-up interval, the wake-up module sends a command to wake up the user at the end of the wake-up interval.

7. The multimodal collaborative wristband wake-up system based on sleep stage perception according to any one of claims 3 and 6, characterized in that, The wake-up time interval is greater than or equal to 30 minutes, and the vital signs monitoring module monitors the sleep stage every 30 seconds.

8. A multimodal collaborative wristband wake-up system based on sleep stage perception according to claim 7, characterized in that, The wake-up module includes a tactile vibration module, a lighting control module, and a bone conduction speaker module. The tactile vibration module includes a linear enhancement segment, a pulse vibration segment, and a random vibration segment; The lighting control module is wirelessly connected to the lighting fixture and controls the lighting fixture to increase from the initial temperature measurement to the preset color temperature during the pulse vibration segment and the random vibration segment.

9. A multimodal collaborative wristband wake-up system based on sleep stage perception according to claim 8, characterized in that, The color temperature change function of the lamp is as follows: In the formula, T ( t () indicates time t The color temperature value; Indicates the initial color temperature; Indicates the target color temperature; k Indicates the slope of the growth rate; This indicates the time at which the color temperature changes.

10. A multimodal collaborative wristband wake-up system based on sleep stage perception according to claim 8, characterized in that, The bone conduction speaker module increases from the initial decibel level to the preset decibel level after the random vibration segment ends; The volume change function of a bone conduction loudspeaker is: ； In the formula, Indicates the initial volume; Indicates the target volume; Indicates the duration of growth.

Images