Sleep assistance composite audio generation method and system based on multi-frequency brain waves and medium

Abstract

The invention discloses a sleep assistance composite audio generation method and system based on multi-frequency brain waves and a medium, and belongs to the technical field of digital audio processing, and the technical scheme is characterized by comprising the following steps: S1, collecting a brain wave signal of a user in real time through a brain wave monitoring module; s2, dynamically selecting a target frequency band combination according to the sleep stage and setting an amplitude change rule; s4, pink noise with the frequency spectrum range of 20-200 Hz is injected into the composite audio, the energy proportion of the pink noise is 5%-15%, and the amplitude is dynamically adjusted according to the environmental sound pressure level SPLenv. The method has the effect that the limitation of traditional sleep assisting audio is solved through the technologies of multi-frequency brain wave synchronization, personalized customization, dynamic adjustment, diversified audio combination and the like.

Description

Technical Field

[0001] The present application relates to the field of digital audio processing technology, and more specifically to a method, system and medium for generating sleep-aiding composite audio based on multi-frequency brain waves. Background Art

[0002] Sleep disorders are a common health problem in modern society. About 30% of people worldwide suffer from insomnia, shallow sleep and other problems. Existing sleep assistance technologies are divided into three categories:

[0003] 1. Drug treatment: Although effective in the short term, long-term use may lead to dependence and side effects;

[0004] 2. Psychotherapy: The effect is significant, but it requires a long time and professional guidance, and is less popular;

[0005] 3. Audio assistance: Although traditional audio assistance technologies (such as white noise and binaural beats) are simple and easy to use, they lack personalization and dynamic adjustment capabilities and cannot accurately match the user's brain wave changes, resulting in limited effectiveness.

[0006] With the cross-integration of brain science and artificial intelligence technology, sleep assistance technology based on electroencephalogram (EEG) has become a research frontier. As a macroscopic manifestation of the electrical activity of brain neuron clusters, the frequency, amplitude and phase changes of brain waves have a strict correspondence with the various stages of the sleep cycle. By collecting EEG signals from brain regions such as the occipital lobe and parietal lobe, the user's sleep stage can be analyzed in real time, and matching audio can be generated accordingly, which can theoretically achieve precise control of sleep quality. However, the current related technologies still have significant technical bottlenecks:

[0007] Binaural beat technology: It relies on two-channel audio stimulation with a fixed frequency difference and can only induce a single-band EEG response. For example, during the transition from early sleep (theta wave-dominated) to wakefulness (alpha wave-dominated), this technology cannot synchronously regulate the synergistic effect of multiple EEG bands, resulting in low induction efficiency. At the same time, due to the lack of a real-time EEG monitoring feedback mechanism, it is impossible to dynamically adjust audio parameters based on individual differences and changes in sleep state.

[0008] Machine learning-based sleep assistance technologies: Existing solutions often use static models to analyze physiological data such as EEG and heart rate. While these can assess sleep quality, their audio generation is limited to fixed-pattern outputs. For example, most systems can only generate basic white noise or simple beat combinations, and are unable to dynamically synthesize multi-frequency composite audio based on EEG phase characteristics, making it difficult to achieve deep synchronization with brain neural oscillations.

[0009] After in-depth research, it was found that the existing technology has three core defects:

[0010] Single-frequency limitations: During natural sleep, the brain coordinates multiple brain regions to produce multi-frequency EEG waves. For example, during slow-wave sleep, delta and theta waves require frequency coupling to achieve memory consolidation. Traditional binaural beat technologies, however, only generate a single difference frequency signal and are unable to simulate this complex neural oscillation pattern.

[0011] Phase conflict: When attempting to superimpose multi-frequency signals to simulate natural EEG, waveform interference occurs due to phase differences between the different frequency waves. For example, when a 4Hz theta wave signal and a 1Hz delta wave signal are superimposed in antiphase, the energy in the target frequency band can be attenuated by 30%-50%, severely weakening the neural stimulation effect.

[0012] Static synthesis models: Existing audio generation algorithms use preset parameter ratios and are unable to adapt to the dynamic changes in brainwave frequency, intensity, and phase during sleep. For example, when transitioning from light sleep to deep sleep, the dominant brainwave frequency shifts from theta waves to delta waves. Static models are unable to adjust audio parameters in real time, resulting in reduced neural synchronization efficiency.

[0013] In view of this, the inventors have designed a new sleep assistance method, system and medium that can monitor the dynamic changes of brain waves in real time and intelligently synthesize multi-frequency composite audio. Summary of the Invention

[0014] To address the shortcomings of the existing technology, the present invention aims to provide a method, system, and medium for generating composite sleep-aid audio based on multi-frequency brainwaves. This method overcomes the limitations of traditional sleep-aid audio through multi-frequency brainwave synchronization, personalized customization, dynamic adjustment, and diverse audio combinations, providing a more scientific, efficient, and personalized sleep-aid solution. Its core innovation lies in achieving precise synchronization between brainwaves and audio, dynamically optimizing audio output, and significantly improving sleep-aid effectiveness and user experience. This represents a significant innovation breakthrough, potentially bringing multiple beneficial effects across multiple social, economic, and technological sectors.

[0015] To achieve the above objectives, the present invention provides the following technical solution: a method for generating a sleep-aiding composite audio based on multi-frequency brain waves, comprising the following steps:

[0016] S1. The EEG monitoring module collects the user's EEG signals in real time, extracts the frequency characteristics of delta waves (0.5-4Hz), theta waves (4-8Hz), and alpha waves (8-12Hz), and identifies the sleep stage.

[0017] S2. Dynamically select the target frequency band combination according to the sleep stage and set the amplitude change rule:

[0018] a. During sleep onset, a combination of alpha waves (8-10 Hz) and theta waves (5-6 Hz) was selected, with the alpha wave amplitude decaying by 15% per minute and the theta wave amplitude increasing exponentially (growth rate k = 0.2 / min);

[0019] b. During light sleep, a combination of theta waves (4-5 Hz) and delta waves (2-3 Hz) is selected, with the amplitude of the theta waves remaining constant and the amplitude of the delta waves increasing by 20% every 5 minutes.

[0020] c. Deep sleep is dominated by delta waves (0.5-1 Hz), with an amplitude of ≥70% and superimposed with 1 / f noise (attenuation slope -6 dB / oct);

[0021] S3. Use Hilbert-Huang transform to extract the instantaneous phase of each frequency band, and dynamically adjust the cross-band phase difference to an absolute value ≤π / 4, so that the total harmonic distortion (THD) of the synthesized waveform is less than 2%;

[0022] S4. Inject pink noise with a spectrum range of 20-200 Hz into the composite audio, with an energy share of 5%-15%, and dynamically adjust the amplitude according to the ambient sound pressure level SPL_env. The formula is:

[0023]

[0024] The present invention is further preferred: the delta wave is generated by a triangle wave, and the fundamental wave formula is:

[0025] y=2|mod(ft,1)-0.5|-1

[0026] , where f is the frequency of the δ wave (0.5-3 Hz); the θ wave and the α wave are generated by a sine wave with an additional random phase shift of ±π / 6.

[0027] The present invention is further preferred: the sleep stage identification step includes: analyzing the frequency, amplitude and zero crossing rate (ZCR) of the brain wave signal through an AI algorithm, wherein: the ZCR during the falling asleep period is less than 600 times / second and the proportion of α wave energy is greater than 50%; the ZCR during the deep sleep period is less than 250 times / second and the proportion of δ wave energy is greater than 65%.

[0028] The present invention is further preferred: the cross-band phase difference constraint condition is specifically: for a high-frequency signal with a frequency of f1 and a low-frequency signal with a frequency of f2, their instantaneous phase angles Φ1 and Φ2 satisfy:

[0029] |φ1-(f1 / f2)·φ2|≤π / 4.

[0030] A sleep-aiding composite audio generation system based on multi-frequency brain waves, comprising:

[0031] Brainwave monitoring module, used to collect brainwave signals and extract delta / theta / alpha wave frequency characteristics;

[0032] Data processing and analysis module, with built-in AI algorithm to identify sleep stages and output stage parameters;

[0033] The composite audio generation module comprises: a baseband generation unit for generating a triangle wave / sine wave baseband signal according to the formula of claim 2;

[0034] a dynamic superposition unit, adjusting the amplitude of each frequency band according to the sleep stage modulation rule described in claim 1;

[0035] Phase optimization unit, which achieves cross-band phase alignment (phase difference ≤π / 4) through Hilbert transform;

[0036] A noise injection unit, generating adaptive pink noise according to the formula of claim 1;

[0037] Dynamic adjustment module to optimize audio parameters based on real-time EEG feedback;

[0038] The storage and transmission module is used to store user EEG data, audio configuration and sleep quality reports.

[0039] The present invention is further preferred: the synthetic signal output by the composite audio generation module must meet the following requirements: the energy proportion of the alpha wave (8-10Hz) during the falling asleep period is greater than 50%; the energy proportion of the delta wave (0.5-3Hz) during the deep sleep period is greater than 65%; and the signal validity is ensured by FFT frequency domain analysis and zero crossing rate (ZCR) time domain verification.

[0040] The present invention is further preferred: the brain wave monitoring module is a wearable EEG headband, which includes a dry electrode sensor and a signal preprocessing unit, which can filter out 50 Hz power frequency noise in real time and extract key frequency characteristics of the original brain wave signal.

[0041] A computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the steps of the multi-frequency brainwave sleep-aiding composite audio generation method according to any one of claims 1 to 4 are implemented.

[0042] In summary, the present invention has the following advantages:

[0043] Through technologies such as multi-frequency brainwave synchronization, personalized customization, dynamic adjustment, and diverse audio combinations, this product overcomes the limitations of traditional sleep-aid audio systems, providing a more scientific, efficient, and personalized sleep-aid solution. Its core innovation lies in the precise synchronization of brainwaves and audio, dynamically optimizing audio output, significantly enhancing sleep-aid effectiveness and user experience. This significant innovation breakthrough promises to bring about multifaceted benefits across social, economic, and technological sectors. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 is a flow chart of the overall method of this embodiment;

[0045] Figure 2 is a flowchart of sleep stage identification in this embodiment;

[0046] Figure 3 is a flow chart of composite audio generation in this embodiment;

[0047] Figure 4 is a diagram of the system module interaction of this embodiment;

[0048] Figure 5 It is a diagram of the internal structure of the audio generation module of this embodiment. DETAILED DESCRIPTION

[0049] The present invention will be further described in detail below with reference to the accompanying drawings.

[0050] A sleep-aiding composite audio generation method based on multi-frequency brain waves, such as Figure 1 、 2 , 3,

[0051] S1. Collect the user's brainwave signals in real time through the brainwave monitoring module, extract the frequency characteristics of delta waves (0.5-4Hz), theta waves (4-8Hz), and alpha waves (8-12Hz), and identify the sleep stage; S11. Feature extraction:

[0052] Energy calculation, calculate the average energy of the signal in each frequency band:

[0053] For a time domain signal x(t), energy is defined as:

[0054]

[0055] Calculate the delta wave energy Eδ, the theta wave energy Eθ, and the alpha wave energy Eα respectively.

[0056] Zero Crossing Rate (ZCR) calculation:

[0057] Calculate how often the signal crosses zero:

[0058]

[0059] Where sign(x) is the sign function, which returns 1 when x ≥ 0, otherwise returns -1. S12. Sleep stage determination:

[0060] Deep sleep period determination, if all the following are met:

[0061] ZCR<250 times / second

[0062]

[0063] It is judged as deep sleep period;

[0064] The sleep period is determined if:

[0065] ZCR<600 times / second

[0066]

[0067] It is determined to be the sleep period;

[0068] Light sleep period determination: If the above conditions are not met, it is determined to be a light sleep period.

[0069] S2. The frequency band combination selection based on sleep stage is as follows:

[0070] S21. The frequency band combination based on the sleep stage is as follows: Drowsy:

[0071] Frequency band selection: Alpha wave (8-10Hz) and theta wave (5-6Hz) combined amplitude dynamic adjustment:

[0072] The alpha wave amplitude decays exponentially:

[0073]

[0074] Where Aα0 = 0.8, t is time (seconds);

[0075] The amplitude of theta waves increases exponentially:

[0076]

[0077] Where Aθ0 = 0.2.

[0078] Light Sleep:

[0079] Frequency band selection: Dynamic adjustment of the amplitude of the combination of theta wave (4-5Hz) and delta wave (2-3Hz):

[0080] Theta wave amplitude is constant: Aθ=0.5

[0081] The delta wave amplitude increases by 20% every 5 minutes:

[0082]

[0083] Where Aδ0=0.3;

[0084] Deep Sleep:

[0085] Frequency band selection: Delta wave (0.5-1Hz) dominant

[0086] Dynamic amplitude adjustment:

[0087] Delta wave amplitude ≥ 70%: Aδ = 0.7

[0088] Superimpose 1 / f noise (attenuation slope -6dB / oct).

[0089] S22, baseband signal generation

[0090] Delta wave generation:

[0091] Using the triangle wave formula:

[0092]

[0093] Where T = 1 / f, f is the frequency of the delta wave (0.5-3Hz)

[0094] Theta and Alpha Wave Generation:

[0095] Using the sine wave formula:

[0096] y(t)=A·sin(2πft+φ)

[0097] Additional ±π / 6 random phase offset: θ∈[-π / 6,π / 6].

[0098] S3, Phase Optimization

[0099] Hilbert transform:

[0100] Perform Hilbert transform on the signal x(t) of each frequency band to obtain the analytical signal

[0101]

[0102] Extract the instantaneous phase:

[0103]

[0104] Phase difference adjustment:

[0105] Calculate the phase difference between adjacent frequency bands:

[0106] Δφ=|φ1(t)-φ2(t)|

[0107] Adjust the phase by digital phase shifter to ensure: Δφ≤π / 4

[0108] The total harmonic distortion (THD) is controlled at <2%.

[0109] S4, pink noise injection

[0110] S41. Noise generation:

[0111] Generate pink noise with a spectrum range of 20-200Hz (power spectral density is inversely proportional to frequency)

[0112] Generated using the Voss-McCartney algorithm, energy accounts for 5-15%

[0113] S42, dynamic adjustment of amplitude:

[0114] Adjust the noise amplitude according to the ambient sound pressure level SPLenv:

[0115]

[0116] S43. Signal synthesis:

[0117] S431, superimposing the optimized signals of each frequency band according to the set amplitude;

[0118] S432, inject adjusted pink noise;

[0119] S433: Output the final composite audio signal.

[0120] The present invention also provides a sleep-aiding composite audio generation system based on multi-frequency brain waves, such as Figure 4 、 5 As shown,

[0121] The EEG monitoring module is a wearable EEG headband. The sensor is a dry electrode (Ag / AgCl material), distributed at FP1, FP2, and O1 positions; the preprocessing circuit is a 50Hz power frequency notch filter (Q=30, attenuation ≥40dB); the bandpass filter group is a δ wave: 0.5-4Hz (3rd order Butterworth, passband ripple ≤0.5dB), θ wave: 4-8Hz, α wave: 8-12Hz, ADC: 16 bits, sampling rate 256Hz, dynamic range 100dB, used to collect EEG signals and extract δ / θ / α wave frequency characteristics.

[0122] The data processing and analysis module can be STM32H743, which implements AI algorithm and can store: 2MB Flash to store historical data, AI algorithm to identify sleep stages and output stage parameters.

[0123] The composite audio generation module generates a triangle wave for the DDS chip (AD9833) with a frequency range of 0.5-3Hz, and theta / alpha wave: the DAC (MCP4922) generates a sine wave with an additional ±π / 6 random phase for dynamic superposition: an analog multiplier (AD633) realizes amplitude modulation, phase optimization: a digital phase shifter (based on FPGA) with an accuracy of π / 32, noise injection, pink noise generation: based on the Voss-McCartney algorithm, spectrum shaping: a 6th-order Butterworth filter realizes -6dB / oct attenuation, and a fundamental frequency generation unit generates a triangle wave / sine wave fundamental frequency signal according to the fundamental wave formula: y=2|mod(ft,1)-0.5|-1.

[0124] The dynamic superposition unit adjusts the amplitude of each frequency band according to the modulation rules of the sleep stage.

[0125] Phase optimization unit, which achieves cross-band phase alignment (phase difference ≤ π / 4) through Hilbert transform; noise injection unit, which generates adaptive pink noise according to the S4 formula;

[0126] The dynamic adjustment module implements a feedback mechanism to re-evaluate the sleep stage and adjust the audio parameters of the PID controller every 30 seconds. The response time is <100ms and the audio parameters are optimized according to the real-time EEG feedback.

[0127] The storage and transmission module is used to store user EEG data, audio configuration and sleep quality reports.

[0128] Furthermore, the synthetic signal output by the composite audio generation module must meet the following requirements: the energy proportion of alpha waves (8-10Hz) during falling asleep is greater than 50%; the energy proportion of delta waves (0.5-3Hz) during deep sleep is greater than 65%; and the signal validity is ensured through FFT frequency domain analysis and zero crossing rate (ZCR) time domain verification.

[0129] Furthermore, the EEG monitoring module is a wearable EEG headband, which includes a dry electrode sensor and a signal preprocessing unit, which can filter out 50Hz power frequency noise in real time and extract the key frequency characteristics of the original EEG signal.

[0130] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of a method for generating a multi-frequency brainwave sleep-aiding composite audio.

[0131] The working process and beneficial effects of the present invention are as follows:

[0132] Through real-time analysis of EEG signals and dynamic modulation of audio parameters, a smart sleep assistance system has been constructed that integrates "acquisition-recognition-generation-feedback." This system transcends the universal applicability limitations of traditional white noise / fixed-frequency audio, enabling precise and personalized sleep intervention. Its technological innovation and proven effectiveness provide a practical solution for smart sleep devices, demonstrating significant technological foresight and market application value.

[0133] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the design concept of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for generating a sleep-aiding composite audio based on multi-frequency brain waves, characterized in that: The following steps are involved: S1. The EEG monitoring module collects the user's EEG signals in real time, extracts the frequency characteristics of delta waves (0.5-4Hz), theta waves (4-8Hz), and alpha waves (8-12Hz), and identifies the sleep stage. S2. Dynamically select the target frequency band combination according to the sleep stage and set the amplitude change rule: a. During sleep onset, a combination of alpha waves (8-10 Hz) and theta waves (5-6 Hz) was selected, with the alpha wave amplitude decaying by 15% per minute and the theta wave amplitude increasing exponentially (growth rate k = 0.2 / min); b. During light sleep, a combination of theta waves (4-5 Hz) and delta waves (2-3 Hz) is selected, with the amplitude of the theta waves remaining constant and the amplitude of the delta waves increasing by 20% every 5 minutes. c. Deep sleep is dominated by delta waves (0.5-1 Hz), with an amplitude of ≥70% and superimposed with 1 / f noise (attenuation slope -6 dB / oct); S3. Use Hilbert-Huang transform to extract the instantaneous phase of each frequency band, and dynamically adjust the cross-band phase difference to an absolute value ≤π / 4, so that the total harmonic distortion (THD) of the synthesized waveform is less than 2%; S4. Inject pink noise with a spectrum range of 20-200 Hz into the composite audio, with an energy share of 5%-15%, and dynamically adjust the amplitude according to the ambient sound pressure level SPL_env. The formula is:

2. The method for generating a sleep-aiding composite audio based on multi-frequency brain waves according to claim 1, characterized in that: The delta wave is generated by a triangle wave, and the fundamental wave formula is: y=2|mod(ft,1)-0.5|-1, Wherein f is the frequency of the δ wave (0.5-3 Hz); the θ wave and the α wave are generated by a sine wave with a random phase shift of ±π / 6 added.

3. The method for generating a sleep-aiding composite audio based on multi-frequency brain waves according to claim 1, characterized in that: The sleep stage identification step includes: analyzing the frequency, amplitude and zero crossing rate (ZCR) of brain wave signals through AI algorithm, wherein: the ZCR during falling asleep is less than 600 times / second and the proportion of alpha wave energy is greater than 50%; the ZCR during deep sleep is less than 250 times / second and the proportion of delta wave energy is greater than 65%.

4. The method for generating a sleep-aiding composite audio based on multi-frequency brain waves according to claim 1, characterized in that: The cross-band phase difference constraint condition is specifically: for a high-band signal with a frequency of f1 and a low-band signal with a frequency of f2, their instantaneous phase angles Φ1 and Φ2 satisfy: |φ1-(f1 / f2)·φ2|≤π / 4.

5. A sleep-aiding composite audio generation system based on multi-frequency brain waves, characterized in that: include: Brainwave monitoring module, used to collect brainwave signals and extract delta / theta / alpha wave frequency characteristics; Data processing and analysis module, with built-in AI algorithm to identify sleep stages and output stage parameters; The composite audio generation module comprises: a baseband generation unit for generating a triangle wave / sine wave baseband signal according to the formula of claim 2; A dynamic superposition unit adjusts the amplitude of each frequency band according to the sleep stage modulation rule of claim 1; a phase optimization unit realizes cross-band phase alignment (phase difference ≤π / 4) through Hilbert transform; A noise injection unit, generating adaptive pink noise according to the formula of claim 1; Dynamic adjustment module to optimize audio parameters based on real-time EEG feedback; The storage and transmission module is used to store user EEG data, audio configuration and sleep quality reports.

6. The sleep-aiding composite audio generation system based on multi-frequency brain waves according to claim 5, characterized in that: The composite audio generation module outputs a synthetic signal that meets the following requirements: the energy proportion of alpha waves (8-10 Hz) during falling asleep is greater than 50%; the energy proportion of delta waves (0.5-3 Hz) during deep sleep is greater than 65%; and the signal validity is ensured through FFT frequency domain analysis and zero crossing rate (ZCR) time domain verification.

7. The sleep-aiding composite audio generation system based on multi-frequency brainwaves according to claim 5, characterized in that: The EEG monitoring module is a wearable EEG headband that includes a dry electrode sensor and a signal preprocessing unit, which can filter out 50Hz power frequency noise in real time and extract the key frequency characteristics of the original EEG signal.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the multi-frequency brainwave sleep-aiding composite audio generation method according to any one of claims 1 to 4 are implemented.

Images