Multi-stage alarm clock awakening system based on biological signal recognition

Abstract

The invention relates to the technical field of intelligent wearable equipment, in particular to a multi-stage alarm clock wake-up system based on biological signal recognition, which collects and fuses multi-modal physiological signals in real time through wearable equipment, dynamically calculates wake-up demand indexes and accurately recognizes sleep stages. Graded combined stimulation of vibration, acoustics and electric tactile sense is adaptively triggered based on the sleep depth, and a mobile terminal remote cooperative control and multi-device networking enhancement mechanism is combined; meanwhile, a gesture detection closed loop is used for verifying the waking state, the stimulation intensity is automatically upgraded when no response is made, and finally progressive waking-up matched with the sleep cycle is achieved. Therefore, the wake-up success rate in the deep sleep period is remarkably improved, the phenomenon of secondary falling asleep is eliminated, and the user experience is improved through collaborative decision while ultra-low power consumption is maintained.

Description

Technical Field

[0001] This invention relates to the field of smart wearable device technology, specifically a multi-level alarm clock wake-up system based on biosignal recognition. Background Technology

[0002] Existing sleep wake-up technologies have many drawbacks. The mechanical wake-up of traditional alarm clocks easily leads users to fall back asleep after turning off the alarm, disrupting the sleep cycle structure and reducing the proportion of deep sleep. According to the Sleep Health Foundation, 38% of users engage in "falling back asleep after turning off the alarm," resulting in a 23% increase in the average lateness rate. The traditional snooze mode's fixed-interval wake-up (usually 9 minutes) disrupts the sleep cycle structure; research by the National Institutes of Health (NIH) shows that it reduces the proportion of deep sleep (N3 stage) by 17%.

[0003] Furthermore, solutions relying solely on ambient sound detection from a mobile phone or vibration from a wristband suffer from high false alarm rates and low success rates in waking up from deep sleep. For example, the false alarm rate for mobile phone-based ambient sound detection is >40% (IEEE Sensors Journal Vol. 21), especially in noisy environments where it cannot distinguish between a user's true awake state. Wristband vibration alone has a wake-up success rate of <15% during deep sleep (N3 stage) (SleepScore Labs test data) and lacks a stimulus intensity regulation mechanism. Simultaneously, while existing wearable devices can collect multimodal biosignals (such as HRV, body movement, and blood oxygenation), they generally lack signal fusion analysis and dynamic linkage with wake-up strategies, resulting in insufficient accuracy in sleep stage determination. These shortcomings collectively contribute to the inefficient wake-up and sleep fragmentation problems of existing sleep wake-up technologies. Summary of the Invention

[0004] The technical problem solved by this invention is to provide a multi-level alarm clock wake-up system based on biosignal recognition to solve the technical problems of inefficient wake-up and fragmented sleep caused by existing sleep wake-up technology.

[0005] The basic solution provided by this invention is a multi-level alarm clock wake-up system based on biosignal recognition, including a wearable device and a mobile terminal. The wearable device includes a biosignal acquisition module, a sleep stage classification module, a multi-level stimulus execution module, and a first collaborative control unit.

[0006] The biosignal acquisition module is used to acquire the user's photoplethysmography (PPG) signal, acceleration signal, and blood oxygen saturation (SpO2) signal.

[0007] The sleep stage classification module is used to calculate the wakefulness demand index (WRI) based on the PPG signal, acceleration signal, and SpO2 signal, and to classify the current sleep stage according to the WRI.

[0008] The multi-level stimulation execution module is used to execute the corresponding combination of vibration stimulation, white noise stimulation and electrotactile stimulation according to the sleep stage classification results.

[0009] The first collaborative control unit is used to communicate with the mobile terminal via the Bluetooth Low Energy (BLE) protocol, receive remote wake-up commands, and feed back physiological data.

[0010] Furthermore, the mobile terminal includes a second collaborative control unit and a user interaction module;

[0011] The second collaborative control unit is used to send a remote wake-up command to the wearable device via the BLE protocol and synchronize the physiological data uploaded by the wearable device;

[0012] The user interaction module is used to receive wake-up requests and timer settings input by the user;

[0013] The wearable device also includes three wake-up process initiation methods: physiological signal triggering, user request triggering, and timed triggering.

[0014] Furthermore, the sleep stage classification module of the wearable device eliminates motion artifacts in the PPG signal by combining an adaptive filtering algorithm with acceleration signals.

[0015] Furthermore, the Wake-Up Demand Index (WRI) is obtained by the wearable device through a weighted fusion calculation based on heart rate variability (HRV), physical activity intensity characteristics, and blood oxygen saturation change trends.

[0016] Furthermore, the sleep stage classification module classifies the current sleep stage according to the WRI, including:

[0017] When 0.4 ≤ WRI < 0.7, it is classified as light sleep; when WRI ≥ 0.7, it is classified as deep sleep; when WRI < 0.4, it is classified as REM sleep.

[0018] Furthermore, the multi-level stimulation execution module of the wearable device operates according to the following rules:

[0019] When classified as light sleep stage, administer 50Hz vibration stimulation and 40-65dB progressively stronger white noise stimulation.

[0020] When classified as deep sleep, 0.5mA of electrical tactile stimulation is superimposed on the vibration and white noise stimulation.

[0021] When classified as REM (Rapid Eye Movement) phase, 30Hz vibrational stimulation is applied.

[0022] Furthermore, the wearable device also includes a gesture detection module;

[0023] The gesture detection module is also used to detect the user's gestures after the stimulus is executed to confirm the awake state; when the wearable device fails to confirm the awake state through the gesture detection module, it automatically increases the stimulation level of the multi-level stimulus execution module after a preset delay time.

[0024] Furthermore, the wearable devices include one or more, and when multiple wearable devices form a collaborative wake-up network through a first collaborative control unit; when the wearable device acting as the master device detects an abnormal sleep state, other wearable devices in the linkage network synchronously increase the stimulation intensity.

[0025] Furthermore, the remote wake-up command of the mobile terminal includes:

[0026] Instant Wake-up Mode: Triggers the wearable device to execute the current optimal stimulation strategy in real time;

[0027] Scheduled wake-up mode: Writes real-time clock (RTC) timed trigger parameters to the wearable device.

[0028] Furthermore, the wearable device maintains a μA-level low-power monitoring mode in the idle state and responds to wake-up trigger events through a hardware interrupt mechanism.

[0029] The principles and advantages of this invention are as follows: This invention collects and fuses multimodal physiological signals in real time through wearable devices, dynamically calculates the wakefulness demand index to accurately identify sleep stages, and adaptively triggers graded combinations of vibration, acoustic, and electrotactile stimulation based on sleep depth. This is combined with remote collaborative control via mobile terminals and a multi-device networking enhancement mechanism. Simultaneously, gesture detection is used to verify the awake state, and the stimulation intensity is automatically upgraded when there is no response, ultimately achieving a gradual wakefulness that matches the sleep cycle. This significantly improves the success rate of waking up during deep sleep, eliminates the phenomenon of falling asleep again, and enhances the user experience through collaborative decision-making while maintaining ultra-low power consumption at the μA level. Attached Figure Description

[0030] Figure 1 This is a logic block diagram of an embodiment of a multi-level alarm clock wake-up system based on biosignal recognition according to the present invention.

[0031] Figure 2 This is an application network diagram of an embodiment of a multi-level alarm clock wake-up system based on biosignal recognition according to the present invention. Detailed Implementation

[0032] The following detailed description illustrates the specific implementation method:

[0033] The specific implementation process is as follows:

[0034] Example 1

[0035] Example 1 is attached. Figure 1As shown, a multi-level alarm clock wake-up system based on biosignal recognition includes a wearable device and a mobile terminal. The wearable device collects the user's physiological signals, preprocesses the signals through its built-in MCU processor, and then synchronizes the user's physiological signals with the mobile terminal via BLE broadcast. When the wearable device detects that the user is in a light sleep state, it first emits a slight vibration; if no awakening response is detected, the vibration gradually intensifies to a combination of audible prompts and electrotactile stimulation. The mobile terminal maintains data synchronization with the wearable device via a wireless communication protocol and can remotely activate the wake-up process when necessary.

[0036] Specifically, the wearable device includes a biosignal acquisition module, a sleep stage classification module, a multi-level stimulus execution module, a gesture detection module, and a first collaborative control unit.

[0037] The biosignal acquisition module is used to acquire the user's photoplethysmography (PPG) signal, acceleration signal, and blood oxygen saturation (SpO2) signal. In this embodiment, the biosignal acquisition module acquires the PPG signal and blood oxygen saturation (SpO2) through a MAX30102 high-sensitivity photoelectric sensor, and transmits the data via I... 2 The C interface transmits raw data at a sampling rate of 125Hz. The body motion intensity is then monitored at a frequency of 100Hz via an integrated Bosch BMI270 triaxial accelerometer, with a motion detection threshold of 5mg.

[0038] The sleep stage classification module calculates the Wakefulness Request Index (WRI) based on the PPG, acceleration, and SpO2 signals, and classifies the current sleep stage according to the WRI. In this embodiment, the wearable device is equipped with a Nordic RF52840 main control chip. After acquiring physiological signals including PPG, acceleration, and SpO2 signals, the main control chip uses an adaptive filtering algorithm to correct the acquired raw PPG signal in real time to eliminate motion artifact interference. Specifically, the system uses body motion signals acquired by a triaxial accelerometer as a motion noise reference to construct a reference signal vector. Based on the Least Mean Square (LMS) criterion, the filter coefficients are dynamically adjusted to separate physiological signals from motion interference. The corrected signal output by the filter retains physiological characteristics such as heart rate and blood oxygenation, while significantly reducing artifacts caused by arm movement, turning over, or slight shaking, thus improving the accuracy of sleep stage classification. This algorithm constructs a motion noise reference signal using acceleration signals and dynamically adjusts the filter coefficients using the LMS criterion, effectively separating physiological signals from motion interference components. Then, multi-dimensional feature analysis is performed on the preprocessed physiological signals. Specifically, firstly, heart rate variability (HRV) features are extracted from the PPG signal, and the standard deviation (SDNN) of adjacent RR intervals is calculated to reflect the activity of the autonomic nervous system. Secondly, based on data collected by a triaxial accelerometer, the moving average variance (MFI) within a set time window (e.g., 10 seconds) is calculated to reflect the user's physical activity intensity and activity status. Next, linear regression analysis is performed on the blood oxygen saturation (SpO2) signal within a minute-level time window to obtain the slope of SpO2 value change (ΔSpO2), used to assess the trend of respiratory and blood oxygen fluctuations. In addition, other auxiliary features, such as PPG waveform amplitude changes, peak heart rate frequency, and acceleration signal energy spectrum, can be extracted to enhance the accuracy of sleep stage determination. Finally, the above feature parameters are processed using a weighted fusion algorithm to obtain the comprehensive wakefulness demand index (WRI).

[0039] WRI=α·HRVstd / HRVnorm+β·MFI / MFImax+γ·ΔSpO2 / ΔSpO2_max,0≤WRI≤1

[0040] In the formula, HRVstd represents the standard deviation of adjacent RR intervals, used to reflect the activity of the autonomic nervous system; HRVnorm is the baseline normalized value; MFI represents the body motion intensity index calculated within a set time window (e.g., 10 seconds), which can be the variance or energy value of the triaxial acceleration signal; MFImax is the normalization coefficient of body motion intensity; ΔSpO2 represents the linear regression slope of blood oxygen saturation within the time window, used to reflect the trend of blood oxygen change; ΔSpO2_max is the normalization coefficient of blood oxygen change; α, β, γ are empirical weighting parameters, in this embodiment α = 0.4, β = 0.35, γ = 0.25, but not limited to these.

[0041] Using the above formula, the WRI value is limited to between 0 and 1, thus corresponding to a preset threshold: when 0.4 ≤ WRI < 0.7, it is determined to be a light sleep stage; when WRI ≥ 0.7, it is determined to be a deep sleep stage; and when WRI < 0.4, it is determined to be a REM sleep stage. This invention, through the above index fusion and classification method, can effectively improve the accuracy of sleep stage identification, thereby providing a basis for subsequent multi-level stimulus execution.

[0042] The multi-level stimulation execution module in this embodiment includes a low-frequency vibration module, a noise module, and an electrical stimulation module. The low-frequency vibration module uses a TDK T4030 vibration motor, supporting 30-100Hz variable frequency vibration, and is driven by a built-in PWM hardware controller. The noise module uses a Knowles SPH0641LM4H microphone, capable of outputting 40-65dB white noise, with a 2dB / s gradation achieved through a DAC. The electrical stimulation module uses a PulseTech LL-01 low-voltage pulse chip, outputting 0-1mA current in 0.1mA steps. The multi-level stimulation execution module automatically matches the optimal stimulation combination based on the real-time calculated WRI value. When WRI < 0.4, it indicates that the user is in a light sleep state, and only the low-frequency vibration module is activated to generate a short pulse vibration of 50Hz (duration 50ms). When 0.4 ≤ WRI < 0.7, it is judged to be a moderate sleep state, and the system simultaneously starts medium-frequency vibration (75Hz, 100ms pulse) and outputs progressive white noise through the noise module (linearly increasing from 40dB to 65dB, with an increase of 2dB / s). When WRI ≥ 0.7, it is identified as a deep sleep state, triggering a high-intensity combination stimulation: including 100Hz continuous vibration, constant 65dB white noise, and 0.8mA microcurrent tactile feedback (pulse width 100μs) emitted by the electrical stimulation module.

[0043] Furthermore, during the stimulation process of the multi-level stimulus execution module, the gesture detection module captures the user's fist-clenching action using a BMI270 triaxial accelerometer: maintaining an acceleration of ±1.5g on the Y-axis for 2 seconds and a variance of <0.1g on the Z-axis. 2The system determines whether the person is awake. If the action is not detected within 10 minutes after stimulation, the system automatically upgrades the stimulation level (e.g., adding electrotactile stimulation during light sleep). Furthermore, the wearable device in this embodiment maintains a μA-level low-power monitoring mode in idle state and responds to wake-up trigger events via a hardware interrupt mechanism. Specifically, the system preprocesses the raw bioelectrical signal using a hardware filtering circuit and employs a hysteresis comparator to eliminate signal jitter interference; when the PPG sensor detects that the heart rate variability exceeds a preset threshold, a hardware interrupt signal is generated. Upon receiving the interrupt signal, the main control chip immediately saves the current operating context and initiates emergency task scheduling in the real-time operating system. Based on the sleep stage assessment results, the system triggers tactile feedback through real-time tasks, uses a PWM hardware accelerator to generate precise vibration waveforms, and dynamically adjusts the execution power.

[0044] The mobile terminal includes a second collaborative control unit and a user interaction module. The first collaborative control unit of the wearable device and the second collaborative control unit of the mobile terminal interconnect via Bluetooth 5.3. The user interaction module receives user-input wake-up requests and timing settings. The second collaborative control unit can send remote wake-up commands to the wearable device via the BLE protocol and synchronize physiological data uploaded by the wearable device. The first collaborative control unit can receive the remote wake-up commands from the second collaborative control unit and provide feedback on physiological data. In this embodiment, the remote wake-up commands of the mobile terminal include an immediate wake-up mode and a scheduled wake-up mode. The immediate wake-up mode triggers the wearable device to execute the current optimal stimulation strategy in real time. The scheduled wake-up mode writes real-time clock (RTC) timing trigger parameters to the wearable device. When the RTC timer reaches a preset value, the main control MCU executes the wake-up command and loads a multi-level stimulation module for wake-up.

[0045] Specifically, in this embodiment, the mobile terminal is a mobile phone, and the wearable device is a smart bracelet. In other embodiments, other types of wearable devices may also be used. The instructions sent by the mobile terminal in this embodiment include an ECDSA digital signature, a 32-byte random number, and an 8-byte timestamp. After successful verification by the wearable device, it executes either immediate wake-up or RTC timed setting. If signature verification fails, a 5-second cooldown period is automatically activated. For example... Figure 2 As shown, when multiple wearable devices are networked together, if the master device detects an SpO2 value below 90% for 10 seconds, it determines it to be abnormal sleep. It then coordinates with surrounding devices via the TDMA relay protocol to synchronously enhance stimulation, improving wake-up efficiency through spatially distributed multi-point stimulation. The system dynamically evaluates the relay suitability of each node in the network. The data forwarding protocol includes: a TDMA frame structure, Reed-Solomon forward error correction, and a packet retransmission mechanism. Furthermore, the master device periodically broadcasts synchronization beacons to support rapid network reconfiguration.

[0046] The above are merely embodiments of the present invention. Commonly known structures and characteristics are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are aware of all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can, under the guidance of this application, improve and implement this solution in combination with their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A multi-stage alarm clock wake-up system based on biological signal recognition, comprising a wearable device and a mobile terminal, characterized in that: The wearable device comprises a biological signal acquisition module, a sleep stage classification module, a multi-level stimulation execution module, and a first collaborative control unit. The biological signal acquisition module is configured to acquire a photoplethysmography (PPG) signal, an acceleration signal, and a blood oxygen saturation (SpO2) signal of a user. The sleep stage classification module is configured to calculate a wake-up demand index (WRI) based on the PPG signal, the acceleration signal, and the SpO2 signal, and classify a current sleep stage according to the WRI. The multi-level stimulation execution module is configured to execute corresponding vibration stimulation, white noise stimulation, and electrohaptics stimulation combination according to the sleep stage classification result. The first collaborative control unit is configured to communicate with a mobile terminal through a Bluetooth Low Energy (BLE) protocol, receive a remote wake-up instruction, and feed back physiological data.

2. The multi-stage alarm clock wake-up system based on biological signal recognition according to claim 1, characterized in that: The mobile terminal comprises a second collaborative control unit and a user interaction module. The second collaborative control unit is configured to send a remote wake-up instruction to the wearable device through the BLE protocol, and synchronize physiological data uploaded by the wearable device. The user interaction module is configured to receive a wake-up request and a timing setting input by a user. The wearable device further comprises three wake-up process starting modes, i.e., a physiological signal triggering mode, a user request triggering mode, and a timing triggering mode.

3. The multi-stage alarm clock wake-up system based on biological signal recognition according to claim 2, characterized in that: The sleep stage classification module of the wearable device eliminates motion artifacts in the PPG signal through an adaptive filtering algorithm combined with the acceleration signal.

4. The multi-stage alarm clock wake-up system based on biological signal recognition of claim 3, wherein: The WRI is calculated by the wearable device based on heart rate variability (HRV), body movement intensity features, and blood oxygen saturation variation trend.

5. The multi-stage alarm clock wake-up system based on biological signal recognition of claim 4, wherein: The sleep stage classification module classifies a current sleep stage according to the WRI, which includes: When 0.4≤WRI<0.7, the sleep stage is classified as light sleep; when WRI≥0.7, the sleep stage is classified as deep sleep; and when WRI<0.4, the sleep stage is classified as rapid eye movement (REM) period.

6. The multi-stage alarm clock wake-up system based on biological signal recognition of claim 5, wherein: The multi-level stimulation execution module of the wearable device operates according to the following rules: When the sleep stage is classified as light sleep, 50Hz vibration stimulation and 40-65dB gradually increasing white noise stimulation are executed; when the sleep stage is classified as deep sleep, 0.5mA electrohaptics stimulation is superimposed on the vibration and white noise stimulation; and when the sleep stage is classified as REM period, 30Hz vibration stimulation is executed. The wearable device further comprises a gesture detection module. The gesture detection module is further configured to detect a user gesture after the stimulation is executed to confirm a wake-up state; and when the wearable device does not confirm the wake-up state through the gesture detection module, the stimulation level of the multi-level stimulation execution module is automatically increased after a preset delay time.

7. The multi-stage alarm clock wake-up system based on biological signal recognition of claim 6, wherein: The wearable device comprises one or more devices, and when multiple wearable devices form a collaborative wake-up network through the first collaborative control unit, if an abnormal sleep state is detected by a wearable device as a master device, the stimulation intensity of other wearable devices in the network is synchronously increased. The remote wake-up instruction of the mobile terminal comprises:

8. The multi-stage alarm clock wake-up system based on biological signal recognition of claim 7, wherein: An immediate wake-up mode, which triggers the wearable device to execute a current optimal stimulation strategy in real time; and a pre-arranged wake-up mode, which writes a real-time clock (RTC) timing trigger parameter to the wearable device.

9. The multi-stage alarm clock wake-up system based on biological signal recognition of claim 8, wherein: The wearable device maintains a μA-level low-power monitoring mode in an idle state, and responds to a wake-up trigger event through a hardware interrupt mechanism. ​ ​ 10. The multi-stage alarm clock wake-up system based on biological signal recognition of claim 9, wherein: ​

Images