Intelligent sleep aiding system and method based on multiband infrared spectrum and physiological feedback

Abstract

The invention discloses an intelligent sleep aiding system and method based on multiband infrared spectrum and physiological feedback. The system comprises a multiband infrared spectrum acquisition module used for acquiring infrared spectrum information of a human body; the physiological feedback signal acquisition module is used for monitoring heart rate, respiratory rate and electroencephalogram signals in real time; the data analyzing and processing module is used for performing fusion analysis on the data by using a deep learning technology so as to identify the sleep state and characteristics of the user; and the sleep-aiding execution module is used for executing personalized sleep-aiding intervention according to the analysis result, and the personalized sleep-aiding intervention comprises music playing and environment temperature and humidity adjustment. According to the method, through collaborative data acquisition, deep analysis processing, sleep demand judgment and accurate execution of sleep aiding measures, accurate monitoring and personalized adjustment of the sleep state are realized, and the sleep quality of the user is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent health technology, specifically to a system and method for achieving intelligent sleep assistance based on the fusion analysis of multi-band infrared spectroscopy and physiological feedback signals, applicable to various scenarios such as daily family life, medical diagnosis and treatment, and care for special populations. Background Technology

[0002] With the fast pace of modern life, sleep disorders have become a significant issue affecting people's physical and mental health. Existing sleep aids, such as simple music players, aromatherapy diffusers, or fitness trackers, are mostly single-function or rely solely on limited physiological signals (such as heart rate) for intervention, lacking a comprehensive and accurate understanding of the user's sleep state, resulting in limited sleep-aiding effects. For example, smart bracelets primarily infer sleep stages through body movement and heart rate variability, but their accuracy needs improvement, and they struggle to proactively and predictively intervene before the user falls asleep. Therefore, there is an urgent need for an intelligent system capable of multi-dimensionally and accurately sensing the user's physiological state and providing personalized, dynamically adjusted sleep-aid strategies accordingly. Summary of the Invention

[0003] Purpose of the invention: The purpose of this invention is to overcome the shortcomings of the prior art and provide an intelligent sleep aid system and method with more accurate monitoring and more personalized intervention.

[0004] Technical Solution: To solve the above-mentioned technical problems, this invention provides an intelligent sleep aid system and method based on multi-band infrared spectroscopy and physiological feedback, as detailed below:

[0005] The intelligent sleep aid system comprises four core modules, which work together to achieve accurate monitoring and personalized sleep assistance:

[0006] Multi-band infrared spectral acquisition module: Equipped with a multi-band infrared sensor, it can cover multiple bands such as near-infrared and mid-infrared, and collect infrared spectral information of the human body surface in a non-contact manner to capture physiological metabolic characteristic data related to sleep.

[0007] Physiological feedback signal acquisition module: integrates heart rate sensor, respiratory rate monitor and EEG signal acquisition electrode to monitor the user's heart rate, respiratory rate fluctuations and EEG signal waveforms in real time and continuously, and obtain core physiological indicators of sleep state.

[0008] Data analysis and processing module: Built-in deep learning model, which integrates and analyzes multi-band infrared spectral data and physiological feedback signals to accurately identify the user's sleep state such as wakefulness, light sleep, and deep sleep, as well as characteristics such as difficulty falling asleep and sleep interruption.

[0009] The sleep aid execution module includes an audio playback unit and a temperature and humidity control unit. It performs personalized sleep aid interventions based on data analysis results, such as playing soothing music during the sleep onset stage and maintaining a suitable temperature and humidity during the deep sleep stage.

[0010] The intelligent sleep method includes the following steps:

[0011] Collaborative data acquisition: After the system is started, the multi-band infrared spectrum acquisition module and the physiological feedback signal acquisition module work synchronously to acquire the user's infrared spectrum information, heart rate, respiratory rate and EEG signals in real time to ensure data integrity.

[0012] Deep analysis and processing: The data analysis and processing module receives the collected multidimensional data, performs feature extraction and fusion calculations through a deep learning model, compares it with the sleep state feature library, and determines the user's current sleep state and potential needs.

[0013] Personalized sleep aid intervention: Based on the analysis results, the sleep aid execution module activates the corresponding sleep aid strategy. For example, for users who have difficulty falling asleep, the ambient temperature and humidity are first adjusted to a comfortable range, and then customized soothing audio is played. When the risk of sleep interruption is detected during sleep, the audio volume or temperature and humidity parameters are automatically adjusted.

[0014] Dynamic optimization and adjustment: The system continuously monitors changes in the user's physiological signals and sleep state, and feeds back to the data analysis and processing module in real time to dynamically adjust the sleep-aid intervention plan to ensure that it meets the user's sleep needs throughout the process.

[0015] Beneficial effects

[0016] (1) More accurate monitoring: Combining multi-band infrared spectrum and multi-dimensional physiological feedback signals, it breaks through the limitations of single signal monitoring, comprehensively captures sleep-related physiological characteristics, and greatly improves the accuracy of sleep state recognition.

[0017] (2) Personalized intervention: Based on deep learning fusion analysis, it accurately matches users' sleep needs and provides customized sleep aid solutions, avoiding the problem of poor effect of generalized intervention.

[0018] (3) Strong dynamic adaptability: Real-time monitoring of changes in sleep status and dynamic adjustment of intervention strategies, covering the entire process of falling asleep, during sleep, and waking up, continuously optimizing the sleep aid effect.

[0019] (4) Wide range of applications: The non-contact data collection design and multi-scenario adaptability can meet the needs of daily family use, medical diagnosis and treatment assistance, and care for special groups such as the elderly and people with insomnia. Attached Figure Description

[0020] 1. Figure 1 System flowchart provided for the implementation of the present invention Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention.

[0022] This embodiment provides an intelligent sleep aid system, which includes:

[0023] Multi-band infrared spectral acquisition module: This module includes multiple infrared light sources (such as near-infrared, mid-infrared, and far-infrared) and corresponding infrared sensors, which are placed in the user's pillow or bedside device. By emitting infrared light of different bands and receiving the spectral signals reflected or transmitted through human tissue, it acquires physiological information related to human blood oxygen saturation, cerebral blood flow, skin temperature, etc.

[0024] Physiological feedback signal acquisition module: This module integrates a heart rate sensor, a respiration sensor, a body movement sensor, and a skin conductance sensor to acquire multidimensional physiological signals such as the user's heart rate variability (HRV), respiratory rate, body movement rate, and skin conductance response.

[0025] Signal Processing and Fusion Analysis Module: This module receives raw signals from the two modules mentioned above, performs preprocessing operations such as filtering, denoising, and feature extraction, and uses machine learning algorithms (such as support vector machines, random forests, or deep learning networks) to fuse and analyze multi-band infrared spectral data and physiological feedback signals to identify the user's sleep stage (such as wakefulness, light sleep, deep sleep, REM sleep) and sleep quality score.

[0026] Intelligent Adjustment Module: Integrates Wi-Fi, Bluetooth, and ZigBee wireless communication modules, supporting communication with devices from mainstream smart home brands (such as Xiaomi, Huawei, and Haier). It features a built-in PID control algorithm that generates control commands based on sleep monitoring data and preset rules, sending them to smart home devices. For example, when indoor humidity is detected to be below 40%, a start command is sent to the humidifier; if the user is in deep sleep and the light intensity is above 5 lux, the curtains are closed and the lights are dimmed.

[0027] Dedicated mobile applications can also be developed, allowing users to interact with the device via a mobile app. The app interface is designed to be simple and intuitive, providing functions such as sleep data visualization charts (e.g., sleep stage time distribution, heart rate change curves), sleep quality scores, and improvement suggestions.

[0028] This embodiment provides an intelligent sleep aid method, which includes:

[0029] S101 signal acquisition stage:

[0030] Non-contact spectral acquisition: A multi-band infrared sensing module deployed above the bedside automatically activates when the user lies down. Its near-infrared band (700-1400nm) is used to detect the periodic fluctuations of subcutaneous blood oxygen saturation (SpO2); the mid-infrared band (3-5μm) is used to monitor the distribution and changes in body surface temperature; and the far-infrared band (8-14μm) senses the intensity of human body thermal radiation, together constituting a spectral feature set of the user's basal metabolism and blood circulation status.

[0031] Contact-based physiological signal acquisition: Synchronous acquisition via high-precision sensors embedded in mattresses or wearable devices.

[0032] ECG signals: used to calculate heart rate variability, focusing on analyzing the ratio of low-frequency power to high-frequency power to assess the state of autonomic nervous system balance; Chest and abdominal respiratory movements: respiratory waveforms are acquired through impedance methods or piezoelectric sensors to calculate respiratory rate, depth, and whether apnea events occur; Body movement signals: the frequency and amplitude of body turning are recorded through a three-dimensional accelerometer to help determine the degree of sleep quietness.

[0033] S102 Signal Processing and Feature Extraction Stage:

[0034] Spectral signal processing: Wavelet transform is performed on the raw infrared spectral data for noise reduction. Then, principal component analysis is used to extract the feature vectors most relevant to sleep state, such as blood oxygen fluctuation entropy value and blood perfusion rate in the frontal lobe region.

[0035] Physiological signal analysis: R-wave detection is performed on electrocardiogram signals to generate RR interval sequences, and then HRV time-domain and frequency-domain indices such as SDNN, RMSSD, and LF / HF are calculated; envelope extraction and spectral analysis are performed on respiratory signals to identify periodic breathing or apnea-hypopnea index; body movement data and HRV data are fused to construct a sleep-wake discriminant function to preliminarily distinguish between wakefulness and sleep states.

[0036] S103 Sleep State Recognition Stage:

[0037] The extracted multimodal features (20+ dimensions in total) are input into a pre-trained lightweight gradient boosting tree model. This model has been trained on a large clinical sleep database and can achieve four-category identification (wakefulness, N1+N2 stage light sleep, N3 stage deep sleep, and REM sleep) with an accuracy of up to 92%. The model outputs a sleep stage result every 30 seconds and, by combining the data from the entire night, generates a quantitative sleep quality index, which is calculated by weighting four dimensions: sleep efficiency, percentage of deep sleep, percentage of REM sleep, and number of awakenings.

[0038] S104 Dynamic Sleep Aid Stage:

[0039] This stage employs a strategy combining state machines and fuzzy control to achieve smooth and human-centered intervention.

[0040] Scenario A: Difficulty falling asleep (awake period exceeding 20 minutes, and HRV showing sympathetic nerve activity), activate the audio-visual synchronization sequence: First, the lights are slowly dimmed to a warm yellow (color temperature 2700K), while playing a binaural beat at a frequency of 8-12Hz to guide the brain to generate alpha waves; if sleep is not achieved after 30 minutes, the system will automatically switch to progressive muscle relaxation voice guidance.

[0041] Scenario B: Maintain deep sleep, turn off all active audio-visual stimulation, and maintain only environmental monitoring. The temperature control system maintains the room temperature within a range of 0.5-1.0℃ lower than when falling asleep, in accordance with the nighttime drop in the body's core temperature.

[0042] Scenario C: During REM sleep, when monitoring revealed typical irregular fluctuations in respiration and heart rate characteristic of REM sleep, extremely low-volume pink noise (<30 dB) was injected. Studies have shown that pink noise helps stabilize sleep structure and reduce nighttime micro-awakenings.

[0043] S105 Feedback Optimization Phase:

[0044] Data pool construction: The system establishes an encrypted personal sleep data pool for each user, storing all data for each sleep cycle, including raw signals, features, identification results, intervention logs, and subjective feedback for the next day.

[0045] Reinforcement learning optimization: Based on these experiences, the system updates its decision-making model using algorithms such as policy gradients. This process enables the system to gradually learn the optimal intervention strategy for a specific user, that is, to select the intervention action that best promotes sleep depth and satisfaction under specific physiological conditions, thereby achieving the evolution from "general intervention" to "precise and personalized sleep aid".

Claims

1. An intelligent sleep aid system based on multi-band infrared spectroscopy and physiological feedback, characterized in that, include: A multi-band infrared spectral acquisition module is used for non-contact acquisition of infrared spectral data from the user's body surface; A physiological feedback signal acquisition module is used to acquire at least one physiological signal from the user, including heart rate, respiratory rate or electroencephalogram (EEG) signal. The data analysis and processing module is used to fuse and analyze the infrared spectral data and physiological feedback signals to identify the user's sleep state; The sleep aid execution module is used to perform personalized sleep aid interventions based on the sleep state.

2. The system according to claim 1, characterized in that, The multi-band infrared spectral acquisition module includes a light source and a sensor for at least two of the near-infrared, mid-infrared, and far-infrared bands.

3. The system according to claim 1, characterized in that, The physiological feedback signal acquisition module includes a heart rate sensor, a respiration sensor, a body movement sensor, or an electroencephalogram (EEG) signal acquisition electrode.

4. The system according to claim 1, characterized in that, The data analysis and processing module uses machine learning or deep learning models to fuse and analyze the infrared spectral data and physiological feedback signals, and outputs sleep stage identification results including wakefulness, light sleep, deep sleep or REM sleep.

5. The system according to claim 1, characterized in that, The sleep aid execution module includes an audio playback unit, an environmental temperature and humidity adjustment unit or a light control unit, and is communicatively connected to smart home devices.

6. A smart sleep aid method based on multi-band infrared spectroscopy and physiological feedback, characterized in that, Includes the following steps: Simultaneously collect the user's infrared spectral data and at least one physiological feedback signal; The infrared spectral data and physiological feedback signals are fused and analyzed to identify the user's current sleep state; Based on the described sleep state, implement corresponding sleep-aiding intervention strategies.

7. The method according to claim 6, characterized in that, The fusion analysis includes: Infrared spectral data were processed to extract features related to sleep states; Process physiological feedback signals to extract heart rate variability, respiratory rate, or body movement characteristics; The extracted features are input into a pre-trained machine learning model, which outputs sleep stage classification results.

8. The method according to claim 6, characterized in that, The sleep-aid intervention strategies include at least one of the following: adjusting the ambient light color temperature, playing audio at a specific frequency, adjusting the ambient temperature and humidity, or playing pink noise.

9. An electronic device comprising a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the method as described in any one of claims 6 to 8.

Images