[go: up one dir, main page]

CN118398169A - Deep learning-based sleep regulation method, system, storage medium and equipment - Google Patents

Deep learning-based sleep regulation method, system, storage medium and equipment Download PDF

Info

Publication number
CN118398169A
CN118398169A CN202311250571.6A CN202311250571A CN118398169A CN 118398169 A CN118398169 A CN 118398169A CN 202311250571 A CN202311250571 A CN 202311250571A CN 118398169 A CN118398169 A CN 118398169A
Authority
CN
China
Prior art keywords
sleep
user
time
sleep state
regulation
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202311250571.6A
Other languages
Chinese (zh)
Inventor
周娜
施一直
毛海央
杨华彬
李俊峰
罗军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Institute of Microelectronics of CAS
Original Assignee
Institute of Microelectronics of CAS
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Institute of Microelectronics of CAS filed Critical Institute of Microelectronics of CAS
Priority to CN202311250571.6A priority Critical patent/CN118398169A/en
Publication of CN118398169A publication Critical patent/CN118398169A/en
Pending legal-status Critical Current

Links

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61M—DEVICES FOR INTRODUCING MEDIA INTO, OR ONTO, THE BODY; DEVICES FOR TRANSDUCING BODY MEDIA OR FOR TAKING MEDIA FROM THE BODY; DEVICES FOR PRODUCING OR ENDING SLEEP OR STUPOR
    • A61M21/00—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis
    • A61M21/02—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis for inducing sleep or relaxation, e.g. by direct nerve stimulation, hypnosis, analgesia
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61M—DEVICES FOR INTRODUCING MEDIA INTO, OR ONTO, THE BODY; DEVICES FOR TRANSDUCING BODY MEDIA OR FOR TAKING MEDIA FROM THE BODY; DEVICES FOR PRODUCING OR ENDING SLEEP OR STUPOR
    • A61M21/00—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
    • G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
    • G16H10/65—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records stored on portable record carriers, e.g. on smartcards, RFID tags or CD
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
    • G16H20/70—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to mental therapies, e.g. psychological therapy or autogenous training
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61M—DEVICES FOR INTRODUCING MEDIA INTO, OR ONTO, THE BODY; DEVICES FOR TRANSDUCING BODY MEDIA OR FOR TAKING MEDIA FROM THE BODY; DEVICES FOR PRODUCING OR ENDING SLEEP OR STUPOR
    • A61M21/00—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis
    • A61M2021/0005—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis by the use of a particular sense, or stimulus
    • A61M2021/0022—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis by the use of a particular sense, or stimulus by the tactile sense, e.g. vibrations
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61M—DEVICES FOR INTRODUCING MEDIA INTO, OR ONTO, THE BODY; DEVICES FOR TRANSDUCING BODY MEDIA OR FOR TAKING MEDIA FROM THE BODY; DEVICES FOR PRODUCING OR ENDING SLEEP OR STUPOR
    • A61M21/00—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis
    • A61M2021/0005—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis by the use of a particular sense, or stimulus
    • A61M2021/0027—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis by the use of a particular sense, or stimulus by the hearing sense
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61M—DEVICES FOR INTRODUCING MEDIA INTO, OR ONTO, THE BODY; DEVICES FOR TRANSDUCING BODY MEDIA OR FOR TAKING MEDIA FROM THE BODY; DEVICES FOR PRODUCING OR ENDING SLEEP OR STUPOR
    • A61M21/00—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis
    • A61M2021/0005—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis by the use of a particular sense, or stimulus
    • A61M2021/0044—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis by the use of a particular sense, or stimulus by the sight sense
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61M—DEVICES FOR INTRODUCING MEDIA INTO, OR ONTO, THE BODY; DEVICES FOR TRANSDUCING BODY MEDIA OR FOR TAKING MEDIA FROM THE BODY; DEVICES FOR PRODUCING OR ENDING SLEEP OR STUPOR
    • A61M21/00—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis
    • A61M2021/0005—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis by the use of a particular sense, or stimulus
    • A61M2021/0083—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis by the use of a particular sense, or stimulus especially for waking up

Landscapes

  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Public Health (AREA)
  • General Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Biomedical Technology (AREA)
  • Medical Informatics (AREA)
  • Psychology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Anesthesiology (AREA)
  • Data Mining & Analysis (AREA)
  • Acoustics & Sound (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Hematology (AREA)
  • Animal Behavior & Ethology (AREA)
  • Theoretical Computer Science (AREA)
  • Veterinary Medicine (AREA)
  • Databases & Information Systems (AREA)
  • Pathology (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • General Physics & Mathematics (AREA)
  • Child & Adolescent Psychology (AREA)
  • Developmental Disabilities (AREA)
  • Hospice & Palliative Care (AREA)
  • Psychiatry (AREA)
  • General Engineering & Computer Science (AREA)
  • Social Psychology (AREA)
  • Computing Systems (AREA)
  • Molecular Biology (AREA)
  • Evolutionary Computation (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Artificial Intelligence (AREA)
  • Pain & Pain Management (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

本公开提供了一种基于深度学习的睡眠调控方法、系统、存储介质及设备，该方法包括：采集用户的当前睡眠指标；将当前睡眠指标输入至深度学习模型，并且将当前睡眠指标存储至用户的个人信息数据库；深度学习模型结合个人信息数据库和用户的入睡时间，以当前睡眠指标作为输入，输出用户的睡眠状态预测曲线；根据睡眠状态预测曲线和用户的预定唤醒时间，生成匹配用户的睡眠调控方案；根据睡眠调控方案对用户的睡眠状态进行基于预设方式的调控。本公开利用深度学习模型预测睡眠状态变化，结合预定唤醒时间自动生成睡眠调控方案，并基于该睡眠调控方案对用户的睡眠状态进行主动调控确保使用者在预定唤醒时间处于浅睡状态，提升用户的睡眠质量。

The present disclosure provides a sleep regulation method, system, storage medium and device based on deep learning, the method comprising: collecting the current sleep index of the user; inputting the current sleep index into the deep learning model, and storing the current sleep index into the user's personal information database; the deep learning model combines the personal information database and the user's sleeping time, takes the current sleep index as input, and outputs the user's sleep state prediction curve; generates a sleep regulation plan that matches the user according to the sleep state prediction curve and the user's scheduled wake-up time; regulates the user's sleep state based on a preset method according to the sleep regulation plan. The present disclosure uses a deep learning model to predict changes in sleep state, automatically generates a sleep regulation plan in combination with the scheduled wake-up time, and actively regulates the user's sleep state based on the sleep regulation plan to ensure that the user is in a light sleep state at the scheduled wake-up time, thereby improving the user's sleep quality.

Description

Deep learning-based sleep regulation method, system, storage medium and equipment
Technical Field
The disclosure relates to the technical field of sleep monitoring, in particular to a sleep regulation method, a sleep regulation system, a sleep regulation storage medium and a sleep regulation device based on deep learning.
Background
During normal human night sleep, non-rapid eye movement (NREM) sleep and Rapid Eye Movement (REM) sleep alternate. The sleeper firstly enters NREM sleep from a wakeful state, starts from a sleep-in period (N1) and enters a light sleep period (N2) after 3-7 minutes; after 10-25 minutes, the sleep enters a deep sleep period (N3-N4), and the deep sleep length varies from a few minutes to one hour according to the personal condition of a sleeper; after the deep sleep is finished, the sleep returns to the N1 phase or the N2 phase; then, a REM sleep period (first REM period lasts for 5-10 minutes) is entered, and one sleep period is completed.
Normal adults will experience 4-6 sleep cycles per night when sleeping. In the latter half of sleep, deep NREM sleep gradually decreases and REM sleep gradually lengthens. If awakened during deep sleep, the sleeper is often perceived as very tired and ill-conditioned; sleep paralysis (a phenomenon commonly known as "ghost press") may occur if the REM wakes up during sleep. Therefore, the optimal wake-up timing is the N1 and N2 phases in NREM sleep.
The existing intelligent alarm clock generally determines a wake-up strategy by monitoring the sleep state of a user and judging the sleep stage of the user or predicting the next sleep stage. However, when the intelligent alarm clock monitors or predicts that the user is in the deep sleep stage at the preset wake-up time, the problem that the user can wake-up in advance or wake-up forcefully can be solved, the former can reduce the whole sleep time, and the latter can interrupt the deep sleep, so that the sleeping quality of the user can be adversely affected.
Disclosure of Invention
An object of the embodiments of the present disclosure is to provide a sleep control method, system, storage medium and device based on deep learning, so as to solve the problem in the prior art that when a user wakes up, the sleeping time of the user is reduced or the deep sleep is interrupted, and the sleeping quality of the user is affected.
The embodiment of the disclosure adopts the following technical scheme: a deep learning-based sleep regulation method, comprising: collecting the current sleep index of a user; inputting the current sleep index into a deep learning model, and storing the current sleep index into a personal information database of the user; the deep learning model combines the personal information database and the sleeping time of the user, takes the current sleeping index as input, and outputs a sleeping state prediction curve of the user; generating a sleep regulation scheme matched with the user according to the sleep state prediction curve and the preset wake-up time of the user; and regulating and controlling the sleep state of the user based on a preset mode according to the sleep regulating and controlling scheme so as to enable the sleep state of the user to be matched with the sleep regulating and controlling scheme.
In some embodiments, the deep learning model, in combination with the personal information database and the time of falling asleep of the user, takes the current sleep index as input, outputs a sleep state prediction curve of the user, comprising: the deep learning model predicts the sleep index of the next moment according to the current sleep index, takes the sleep index of the next moment as a new current sleep index, and repeatedly predicts the sleep index of the next moment; and outputting a sleep state prediction curve of the user according to the personal information database and all sleep indexes of the next moment predicted by the deep learning model.
In some embodiments, the generating a sleep regulation scheme that matches the user according to the sleep state prediction curve and the predetermined wake time of the user includes: detecting whether the sleep state corresponding to the preset awakening time in the sleep state prediction curve is a light sleep state or not; under the condition that the sleep state corresponding to the preset awakening time is a light sleep state, the sleep regulation scheme is empty; determining a light sleep state area nearest to the preset awakening time on a time axis of the sleep state prediction curve under the condition that the sleep state corresponding to the preset awakening time is not the light sleep state; and determining the sleep regulation scheme to accelerate or decelerate the sleep cycle of the user according to the sequence between the light sleep state area and the preset wake-up time.
In some embodiments, after the adjusting the sleep state of the user according to the sleep adjusting scheme based on the preset manner, the matching the sleep state of the user with the sleep adjusting scheme further includes: collecting current sleep indexes in real time; determining a current actual sleep state of a user according to the current sleep index, updating the deep learning model according to the actual sleep state, and storing the current sleep index into the personal information database; re-predicting the sleep state prediction curve according to the updated deep learning model; and regenerating a sleep regulation scheme according to the re-predicted sleep state prediction curve and the preset wake-up time of the user, and regulating the sleep state of the user based on a preset mode based on the regenerated sleep regulation scheme.
In some embodiments, further comprising: detecting whether sleep interruption occurs; re-determining the time to fall asleep of the user in the event of sleep disruption; the deep learning model is combined with the personal information database and the redetermined sleeping time of the user to re-predict the sleeping state prediction curve; and regenerating a sleep regulation scheme according to the re-predicted sleep state prediction curve and the preset wake-up time of the user, and regulating the sleep state of the user based on a preset mode based on the regenerated sleep regulation scheme.
In some embodiments, the generating a sleep regulation scheme that matches the user according to the sleep state prediction curve and the predetermined wake time of the user includes: obtaining a dream preference parameter k and a sleep regulation limit time N of a user, wherein k is [ -1,1]; determining a sleep regulation limit M, wherein M= |kN|; determining a first sleep period conforming to the user's dream preference according to the dream preference parameter k, the predetermined wake-up time and the sleep state prediction curve; detecting whether the difference between the first sleep period and the sleep period which is not regulated by the user is smaller than or equal to the sleep regulation limit; under the condition that the difference value between the first sleep period and the sleep period which is not regulated by the user is smaller than or equal to the sleep regulation limit, the sleep regulation scheme is to accelerate or decelerate the sleep period of the user, so that the sleep period coincides with the first sleep period after regulation; and under the condition that the difference between the first sleep period and the sleep period which is not regulated by the user is larger than the sleep regulation limit, determining a light sleep state area closest to the preset awakening time on a time axis of the sleep state prediction curve, and determining the sleep regulation scheme to accelerate or decelerate the sleep period of the user according to the sequence between the light sleep state area and the preset awakening time.
In some embodiments, in a case where the deep learning model cannot output the sleep state prediction curve and/or after the adjustment based on the preset manner, the method further includes: acquiring an elastic awakening interval set by the user, wherein the preset awakening time is positioned in the elastic awakening interval; determining whether a sleep cycle of the user has a light sleep state within the elastic wake-up interval; waking up a user in a time period corresponding to a light sleep state under the condition that the sleep period of the user has the light sleep state; and under the condition that the sleep period of the user does not have a light sleep state, regulating and controlling the user based on a preset mode before the user enters the deep sleep state in the last sleep period of the sleep period in which the preset wake-up time is located so as to prevent the user from entering the deep sleep state, or waking up the user at the time corresponding to the second endpoint of the elastic wake-up interval, wherein the time corresponding to the first endpoint of the elastic wake-up interval is earlier than the time corresponding to the second endpoint.
The embodiment of the disclosure also provides a sleep regulation system based on deep learning, comprising: sleep monitoring system, operation storage system and sensory stimulation system; the sleep monitoring system is used for collecting current sleep indexes of a user; the operation storage system is used for inputting the current sleep index into a deep learning model and storing the current sleep index into a personal information database of the user; the deep learning model combines the personal information database and the sleeping time of the user, takes the current sleeping index as input, and outputs a sleeping state prediction curve of the user; generating a sleep regulation scheme matched with the user according to the sleep state prediction curve and the preset wake-up time of the user; the sensory stimulation system is used for regulating and controlling the sleep state of the user based on a preset mode according to the sleep regulating and controlling scheme so that the sleep state of the user is matched with the sleep regulating and controlling scheme.
The embodiment of the disclosure also provides a storage medium storing a computer program which, when executed by a processor, realizes the steps of the sleep regulation method based on deep learning.
The embodiment of the disclosure also provides equipment, which at least comprises a memory and a processor, wherein the memory is stored with a computer program, and the processor realizes the steps of the sleep regulation method based on deep learning when executing the computer program on the memory.
The beneficial effects of the embodiment of the disclosure are that: and in the sleeping process of the user, the change of the sleeping state is judged and predicted by using a deep learning model, a matched sleeping regulation scheme is automatically generated for the user in combination with the preset awakening time set by the user, and the sleeping state of the user is actively regulated and controlled based on the sleeping regulation scheme to be matched with the sleeping regulation scheme, so that the sleeping quality of the user is improved under the condition that the sleeping time of the user is not influenced by the sleeping state of the user, and the sleeping state of the user is ensured to be in a shallow sleeping state at the preset awakening time.
Drawings
For a clearer description of one or more embodiments of the present description or of the solutions of the prior art, the drawings that are necessary for the description of the embodiments or of the prior art will be briefly described, it being apparent that the drawings in the description that follow are only some of the embodiments described in the description, from which, for a person skilled in the art, other drawings can be obtained without inventive faculty.
FIG. 1 is a flow chart of a deep learning based sleep regulation method in a first embodiment of the present disclosure;
FIG. 2 is a schematic diagram of a sleep regulation scheme in a first embodiment of the present disclosure;
FIG. 3 is a schematic diagram of a sleep regulation scheme based on dream preference in a first embodiment of the present disclosure;
Fig. 4 is a schematic structural diagram of a sleep control device based on deep learning in a second embodiment of the present disclosure.
Detailed Description
In order to enable a person skilled in the art to better understand the technical solutions in one or more embodiments of the present specification, the technical solutions in one or more embodiments of the present specification will be clearly and completely described below with reference to the drawings in one or more embodiments of the present specification, and it is obvious that the described embodiments are only some embodiments of the present specification, not all embodiments. All other embodiments, which can be made by one or more embodiments of the present disclosure without inventive faculty, are intended to be within the scope of the present disclosure.
During normal human night sleep, non-rapid eye movement (NREM) sleep and Rapid Eye Movement (REM) sleep alternate. The sleeper firstly enters NREM sleep from a wakeful state, starts from a sleep-in period (N1) and enters a light sleep period (N2) after 3-7 minutes; after 10-25 minutes, the sleep enters a deep sleep period (N3-N4), and the deep sleep length varies from a few minutes to one hour according to the personal condition of a sleeper; after the deep sleep is finished, the sleep returns to the N1 phase or the N2 phase; then, a REM sleep period (first REM period lasts for 5-10 minutes) is entered, and one sleep period is completed.
Normal adults will experience 4-6 sleep cycles per night when sleeping. In the latter half of sleep, deep NREM sleep gradually decreases and REM sleep gradually lengthens. If awakened during deep sleep, the sleeper is often perceived as very tired and ill-conditioned; sleep paralysis (a phenomenon commonly known as "ghost press") may occur if the REM wakes up during sleep. Therefore, the optimal wake-up timing is the N1 and N2 phases in NREM sleep.
The existing intelligent alarm clock generally determines a wake-up strategy by monitoring the sleep state of a user and judging the sleep stage of the user or predicting the next sleep stage. However, when the intelligent alarm clock monitors or predicts that the user is in the deep sleep stage at the preset wake-up time, the problem that the user can wake-up in advance or wake-up forcefully can be solved, the former can reduce the whole sleep time, and the latter can interrupt the deep sleep, so that the sleeping quality of the user can be adversely affected.
In order to solve the above-mentioned problems, a first embodiment of the present disclosure provides a sleep control method based on deep learning, which can be applied to any intelligent device with sleep monitoring function, such as a smart speaker, a tablet computer, and even a mobile phone. Fig. 1 shows a flowchart of a deep learning-based sleep regulation method in the present embodiment, which mainly includes steps S10 to S50:
S10, collecting the current sleep index of the user.
The intelligent device executing the method of the embodiment can have the sleep index acquisition function of the user, can also communicate with other devices such as acquisition devices or sensors with the sleep index acquisition function, and can wear a series of devices with the sleep index acquisition function before the user is ready to sleep or start the devices with the sleep index acquisition function, and the devices can acquire the current sleep index of the user in real time and serve as data support for subsequent sleep regulation.
The sleep index mainly collected in this embodiment may be any one or more of the following indexes: the equipment for acquiring the sleep indexes can be wearable equipment such as a sports watch and a bracelet, can also be monitoring equipment such as ultrasonic radar placed at the bedside or a sleep monitoring belt placed on the bed, and can also be professional electroencephalogram or electrocardiograph monitoring instruments and the like. In the actual use process, the sleep index collection device of the user can be matched, one or more of the sleep indexes are selected for collection, and the more the types of the sleep indexes are, the better the corresponding monitoring and regulating effects are.
It should be noted that, the collection period of the current sleep index of the user may be adjusted according to the user setting, for example, ten times per second or one time per second, the higher the collection frequency corresponds to the more sleep indexes, the higher the accuracy in the subsequent deep learning model update is, but more energy consumption is also caused; in actual use, the method can also carry out adaptive adjustment according to the sleep state of the user, for example, the sleep index acquisition is carried out at shorter intervals before the user falls asleep, after the user is judged to fall asleep, the acquisition period is adaptively adjusted according to the change condition of the sleep period, so that the acquisition times are reduced as much as possible under the condition that the accurate index under each sleep state is accurately acquired, and the purpose of saving energy is achieved.
S20, inputting the current sleep index into the deep learning model, and storing the current sleep index into a personal information database of the user.
The deep learning model in this embodiment may be a long-short term memory (LTSM) network, where the collected sleep index of the user is used as an input vector to determine the current sleep state of the user, and predict the subsequent sleep state. The personal information database is used for storing the sleep index and the actual sleep state actually collected by the current user and the matching condition between the sleep index and the actual sleep state, and simultaneously records the sleep state change, the duration, the sleep time, the awake time and other data information of the user.
It should be noted that, the data recorded in the personal information database is increased along with the use process of the user, and the recorded data are the actual sleep index and sleep state of the user; the deep learning model is mainly used for prediction, and can be updated and optimized continuously in the subsequent use process, so that the prediction result is more accurate.
S30, the deep learning model combines the personal information database and the sleeping time of the user, takes the current sleeping index as input, and outputs a sleeping state prediction curve of the user.
The deep learning model can determine the current sleep state of the user according to the current sleep index of the user, for example, determine the sleeping time of the user, and predict the subsequent sleep index of the user by combining the acquired sleep index based on the sleeping time.
In some embodiments, the deep learning model predicts the sleep index at the next time based on the current sleep index as the input vector, where the next time refers to a time with a certain time interval from the time when the sleep index is currently collected, and the specific value of the time interval can be adjusted according to the requirement. Further, the deep learning model takes the sleep index at the next moment as a new current sleep index, repeatedly predicts the sleep index at the next moment to obtain a long-term sleep index change prediction, combines the sleep states of the user actually recorded by the personal information database based on all the sleep indexes at the next moment obtained in the long-term prediction, and correspondingly outputs a sleep state prediction curve of the user to represent the possible corresponding sleep state change condition of the user in the current sleep.
S40, generating a sleep regulation scheme matched with the user according to the sleep state prediction curve and the preset wake-up time of the user.
After the sleep state prediction curve is obtained, the sleep state corresponding to the preset wake-up time can be determined according to the preset wake-up time preset by the user, and the sleep regulation scheme is correspondingly generated according to the sleep state corresponding to the preset wake-up time.
Specifically, since the user is awakened in the light sleep state without tired feeling, whether the sleep state corresponding to the preset awakening time is the light sleep state or not is firstly determined in the sleep state prediction curve, and the sleep regulation scheme can be set to be empty under the condition that the sleep state corresponding to the preset awakening time is the light sleep state, namely the sleep state of the user is not regulated, and the user can just be in the light sleep state and just awaken when the preset awakening time arrives according to the normal sleep state change of the user; under the condition that the sleep state corresponding to the preset awakening time is not the light sleep state, if the sleep state is not regulated, the user is possibly awakened in the deep sleep state to influence the sleep quality, at the moment, the sleep state of the user is required to be regulated and controlled through a sleep regulation scheme, namely, a light sleep state area closest to the preset awakening time is determined on a time axis of a sleep state prediction curve, and according to the sequence between the light sleep state area and the preset awakening time, the sleep regulation scheme is determined to accelerate or decelerate the sleep period of the user, so that the sleep period of the user is correspondingly shortened or prolonged, and the sleep state corresponding to the preset awakening time is just in the light sleep state area, so that the user is awakened when the user is in the light sleep state.
S50, regulating and controlling the sleep state of the user based on a preset mode according to the sleep regulating and controlling scheme, so that the sleep state of the user is matched with the sleep regulating and controlling scheme.
The sleep regulation scheme determines a processing scheme for accelerating or decelerating the sleep state of the user according to the corresponding situation of the sleep state prediction curve and the preset wake-up time, and then the sleep state of the user can be regulated and controlled in a preset mode, so that the sleep state of the user is matched with the sleep regulation scheme, the user is in a light sleep state just when the user is waken up in the preset wake-up time, and the sleep quality of the user is prevented from being reduced when the user is waken up in the deep sleep state.
Fig. 2 shows a schematic diagram of a sleep regulation scheme. As shown in fig. 2, the horizontal axis represents sleep time, three curves from top to bottom respectively represent the change condition of sleep state with the lapse of time, three vertical dotted lines represent different preset wake-up times, the area between two horizontal dotted lines in each curve represents light sleep state, and the line segment part of the curve falling into the area between the dotted lines represents the time corresponding to the line segment is in light sleep state. The first curve in fig. 2 shows a sleep state prediction curve, which means normal sleep without sensory stimulation, where the curve positions corresponding to three different predetermined wake-up times do not fall within the range of the light sleep state, and the second curve and the third curve respectively show a curve of slowing down and prolonging the sleep period and a curve of accelerating and shortening the sleep period, where the curve positions corresponding to the adjusted predetermined wake-up times just fall within the region of the light sleep state. It should be noted that in actual use, the predetermined wake-up time set by the user is typically only one, so that only a single predetermined wake-up time is needed for adjustment, and the three different predetermined wake-up times shown in fig. 2 are only used as illustrations.
In this embodiment, the preset modes at least include one or more of the following modes: the regulation and control modes adopted in the specific implementation of playing audio, adjusting ambient light, transcranial electric stimulation and transcranial magnetic stimulation can be adjusted according to the requirements of users or the conditions of equipment of the users, the embodiment is not limited herein, and the regulation and control modes are all realized through corresponding stimulation units in the sensory stimulation system.
It should be noted that, because the sleep states of each person are different, the response to the sensory stimulation may be different, the stimulus intensity setting may be performed based on the pre-training model for most people when the sensory stimulation is performed for the first time, when the user is in the process of or after the regulation, the current sleep index of the user may be further collected in real time, and the current actual sleep state of the user may be determined according to the current sleep index, so as to reflect the regulation effect of the sensory stimulation in real time, and the sensory stimulation intensity may be adjusted according to the deviation between the actual stimulation condition and the predicted stimulation condition.
In some embodiments, the current actual sleep state of the user can be determined by collecting the current sleep index in real time, the deep learning model is updated, and the actual sleep index of the user is stored in the personal information database; when the deep learning model is updated, the weight matrix in the deep learning model is mainly updated, so that the prediction of the sleep index at the next moment and the determination of the corresponding sleep state correspondingly are more in line with the sleeping habit of the current user, and the aim of improving the accuracy is fulfilled. And after updating the deep learning model, predicting a sleep state curve again by utilizing the new deep learning module, further regenerating a sleep regulation scheme, and regulating the sleep state of the user based on a preset mode based on the regenerated sleep regulation scheme when regulating is carried out subsequently.
According to the embodiment, through monitoring of the sleep index in the sleeping process of the user, the change of the sleep state of the user is judged and predicted by using the deep learning model, a matched sleep regulation scheme is automatically generated for the user in combination with the preset wake-up time set by the user, and the sleep state of the user is actively regulated and controlled based on the sleep regulation scheme to be matched with the sleep regulation scheme, so that the user is ensured to be in a shallow sleep state at the preset wake-up time, and the sleeping quality of the user is improved under the condition that the sleeping duration of the user is not influenced.
In addition, in some embodiments, the intelligent device may also detect in real time whether the user has sleep interruption during the sleep process of the user, for example, a nightfall, an emergency is awakened, etc., where the user has a larger sleep index fluctuation or body movement, and the sleep period is also interrupted; when the sleep interruption occurs to the user, the sleeping time of the user is redetermined based on the sleep index acquired in real time, the deep learning model is utilized to combine the personal information database and the redetermined sleeping time to conduct the redevelopment of the sleep state prediction curve, the sleep regulation scheme is regenerated according to the redevelopment of the sleep state prediction curve and the predetermined wake-up time of the user, and the sleep state of the user is regulated based on the regenerated sleep regulation scheme in a preset mode.
In some embodiments, the user may personalize the content of the sleep regulation scheme and its implementation based on his own preferences for dreams. Specifically, a user may preset a dream preference parameter k to represent the self preference condition of the user for the dream content, where k e [ -1,1], when the value of k is closer to 1, indicates that the user desires to learn the dream condition, whereas when the value of k is closer to-1, indicates that the user does not desire to learn the dream condition; in addition, according to the sense organ stimulation mode and the response degree of the user to the sense organ stimulation, a sleep regulation limit duration N is set, the unit is minutes, the difference value between the regulated sleep period of the user and the original sleep period length when not regulated is represented, and the value can be automatically updated through personal data collection of the user in long-term use.
In the process of generating the sleep regulation scheme matched with the user, personalized sleep regulation scheme generation can be performed according to the preference condition of the user for the dream. The dream usually occurs in the REM sleep period of the user, the user can clearly remember the dream content when the REM later in the shallow sleep period is awakened, and the user usually forgets the dream content in the previous REM sleep period when the REM earlier in the shallow sleep period is awakened, so that when the user is actually awakened, the sleep regulation scheme is correspondingly generated according to the preference of the user on whether the dream content is expected to be known or not and the actual effect of sleep regulation of the system.
Firstly, obtaining a dream preference parameter k and a sleep regulation limit duration N of a user, and determining a sleep regulation limit M according to k and N, wherein M= |kN|, and setting the sleep regulation limit to reduce the influence of sleep regulation on the natural sleep of the user as much as possible, wherein in the process of generating a subsequent sleep regulation scheme, equipment performs sleep regulation by taking the sleep regulation limit as a limit; combining the dream preference parameter k, the preset awakening time and the sleep state prediction curve, namely selecting a light sleep state area conforming to the dream preference on the sleep state prediction curve, and determining a first sleep period conforming to the dream preference of the user according to the light sleep state area, namely if the user is awakened in a specific light sleep state according to the dream preference of the user, the sleep period of the user needs to be prolonged or shortened to be the first sleep period, and when the preset awakening time arrives, the user is just in the light sleep state; then detecting whether the difference between the first sleep period and the sleep period which is not regulated by the user is smaller than or equal to the sleep regulation limit, and under the condition that the difference between the first sleep period and the sleep period which is not regulated by the user is smaller than or equal to the sleep regulation limit, namely, the personalized sleep regulation can be realized under the condition that the natural sleep of the user is not excessively influenced, wherein the generated sleep regulation scheme is to accelerate or decelerate the sleep period of the user, so that the sleep period coincides with the first sleep period after regulation; under the condition that the difference between the first sleep period and the sleep period which is not regulated by the user is larger than the sleep regulation limit, namely, if the system regulates according to the dreaminess of the user, natural sleep of the user can be influenced, at the moment, the preference condition of the user to the dreaminess can be not considered, according to the principle of minimum regulation, a light sleep state area closest to the preset awakening time is determined on a time axis of a sleep state prediction curve, and according to the sequence between the light sleep state area and the preset awakening time, the sleep regulation scheme is determined to accelerate or decelerate the sleep period of the user, so that the risk that the sleep of the user is interrupted by the system stimulation or uncomfortable feeling is reduced.
It should be noted that, the user's preference condition of the dream may be reflected according to the value of the dream preference parameter k set by the user, that is, when k is greater than 0, the user is determined to want to learn the dream condition, at this time, the light sleep state area conforming to the dream preference may be determined to be the light sleep stage after REM closest to the predetermined wake-up time, and if k is less than or equal to 0, the user is determined to not want to learn the dream condition, at this time, the light sleep state area conforming to the dream preference may be determined to be the light sleep stage before REM closest to the predetermined wake-up time. If the dream preference of the user is not considered, the light sleep state closest to the preset time is taken into consideration.
FIG. 3 is a schematic diagram of a sleep regulation scheme based on dream preference, as shown in FIG. 3, in which a user is in REM sleep state normally at a predetermined wake-up time, if the user does not want to know the dream, regulating each period with M as a limit, and making the wake-up time be in shallow sleep region before REM by prolonging the sleep period; if the user wants to know the dream condition, the user takes M as a limit to regulate each period, and the sleep period is shortened to enable the wake-up time to be in a shallow sleep area after REM.
In the process of user sleep regulation, when the sleep state cannot be accurately regulated due to insufficient personal sleep data collection of the user or sleep disturbance of the user, the user can be individually awakened according to the awakening preference setting performed in advance by the user. Specifically, the user sets an elastic wake-up interval according to personal requirements, that is, a maximum value of the wake-up time that can be accepted by the user in advance or later than the preset wake-up time, for example, the preset wake-up time of the user is 7 hours, and the elastic wake-up time can be set to be 6 hours 50 minutes to 7 hours 20 minutes, where the two times are respectively a first endpoint and a second endpoint of the elastic wake-up interval, and the embodiment limits the time corresponding to the second endpoint to be later than the time corresponding to the first endpoint; under the condition that the sleep state of the user cannot be regulated and controlled, whether the sleep period of the user in the elastic awakening interval comprises a light sleep state or not can be determined according to a sleep state prediction curve output by a personal information database or a deep learning model of the user, if the sleep period of the user in the elastic awakening interval comprises the light sleep state, the user is awakened in a time period corresponding to the light sleep state in the elastic awakening interval, if the sleep period of the user in the elastic awakening interval does not comprise the light sleep state, the user can be forcibly awakened in a time corresponding to a second endpoint of the elastic awakening interval according to a preset awakening preference of the user, or the regulation and control based on a preset mode is carried out on the user before the user enters the deep sleep state in the last sleep period of the sleep period in which the preset awakening time is positioned, so that the user is prevented from entering the deep sleep state, and the shallow sleep state is ensured at the expense of the deep sleep time.
Based on the same inventive concept, a second embodiment of the present disclosure provides a sleep regulation system based on deep learning, and a structural schematic diagram of the system is shown in fig. 4, and the system at least comprises a sleep monitoring system 10, an operation storage system 20 and a sensory stimulation system 30, wherein the sleep monitoring system 10 is used for collecting current sleep indexes of a user; the operation storage system 20 is used for inputting the current sleep index into the deep learning model and storing the current sleep index into a personal information database of the user; the deep learning model combines the personal information database and the sleeping time of the user, takes the current sleeping index as input, and outputs a sleeping state prediction curve of the user; generating a sleep regulation scheme matched with the user according to the sleep state prediction curve and the preset wake-up time of the user; the sensory stimulation system 30 is used for regulating and controlling the sleep state of the user according to the sleep regulation scheme based on a preset mode, so that the sleep state of the user is matched with the sleep regulation scheme.
Specifically, the sleep index mainly collected in this embodiment may be any one or more of the following indexes: respiratory rate, blood oxygen concentration, heart rate, electroencephalogram index, electrocardiograph index and the like, and the sleep monitoring system 10 can comprise wearable equipment such as a sports watch and a bracelet, or monitoring equipment such as ultrasonic radar placed at the bedside or a sleep monitoring belt placed on the bed, and can also be a professional electroencephalogram or electrocardiograph monitoring instrument.
Specifically, the preset modes performed by the sensory stimulation system 30 include at least one or more of the following modes: playing audio, adjusting ambient light, transcranial electrical stimulation, transcranial magnetic stimulation, the sensory stimulation system 30 may be correspondingly configured as a sound box with audio playing, a light fixture with brightness and frequency adjusting functions, a stimulation device with transcranial electrical stimulation or transcranial magnetic stimulation, and the like.
In some embodiments, the operation storage system 20 is specifically configured to predict a sleep index at a next time according to the current sleep index by using the deep learning model, and repeat the prediction of the sleep index at the next time by using the sleep index at the next time as a new current sleep index; and outputting a sleep state prediction curve of the user according to the sleep indexes of all the next moments predicted by the personal information database and the deep learning model.
In some embodiments, the operation storage system 20 is specifically configured to detect whether a sleep state corresponding to a predetermined wake-up time in the sleep state prediction curve is a light sleep state; under the condition that the sleep state corresponding to the preset awakening time is a light sleep state, the sleep regulation scheme is empty; under the condition that the sleep state corresponding to the preset awakening time is not the light sleep state, determining a light sleep state area nearest to the preset awakening time on a time axis of a sleep state prediction curve; and determining a sleep regulation scheme to accelerate or decelerate the sleep cycle of the user according to the sequence between the light sleep state area and the preset wake-up time.
In some embodiments, the sleep monitoring system 10 is also used to acquire current sleep metrics in real-time; the operation storage system 20 determines the current actual sleep state of the user according to the current sleep index, updates the deep learning model according to the actual sleep state, and stores the current sleep index into the personal information database; re-predicting a sleep state prediction curve according to the updated deep learning model; and regenerating the sleep regulation scheme according to the re-predicted sleep state prediction curve and the preset wake-up time of the user, so that the sensory stimulation system 30 regulates the sleep state of the user based on the regenerated sleep regulation scheme in a preset mode.
In some embodiments, sleep monitoring system 10 is also used to detect whether sleep disruption has occurred; in the event of a sleep disruption, the computing storage system 20 re-determines the user's fall asleep time; the deep learning model is combined with the personal information database and the redetermined sleeping time of the user to re-predict the sleeping state prediction curve; and regenerating the sleep regulation scheme according to the re-predicted sleep state prediction curve and the preset wake-up time of the user, so that the sensory stimulation system 30 regulates the sleep state of the user based on the regenerated sleep regulation scheme in a preset mode.
In some embodiments, the operational storage system 20 is further configured to: obtaining a dream preference parameter k and a sleep regulation limit time N of a user, wherein k is [ -1,1]; determining a sleep regulation limit M, wherein M= |kN|; determining a first sleep period conforming to the user's dream preference according to the dream preference parameter k, the predetermined wake-up time and the sleep state prediction curve; detecting whether the difference between the first sleep period and the sleep period which is not regulated by the user is smaller than or equal to the sleep regulation limit; under the condition that the difference value between the first sleep period and the sleep period which is not regulated by the user is smaller than or equal to the sleep regulation limit, the sleep regulation scheme is to accelerate or decelerate the sleep period of the user, so that the sleep period coincides with the first sleep period after regulation; and under the condition that the difference between the first sleep period and the sleep period which is not regulated by the user is larger than the sleep regulation limit, determining a light sleep state area closest to the preset awakening time on a time axis of the sleep state prediction curve, and determining the sleep regulation scheme to accelerate or decelerate the sleep period of the user according to the sequence between the light sleep state area and the preset awakening time.
In some embodiments, the operational storage system 20 is further configured to: acquiring an elastic awakening interval set by the user, wherein the preset awakening time is positioned in the elastic awakening interval; determining whether a sleep cycle of the user includes a light sleep state within the elastic wake-up interval; when the sleep cycle of the user comprises a light sleep state, waking up the user in a time period corresponding to the light sleep state; and under the condition that the sleep period of the user does not comprise a light sleep state, regulating and controlling the user based on a preset mode before the user enters the deep sleep state in the last sleep period of the sleep period in which the preset wake-up time is positioned so as to prevent the user from entering the deep sleep state, or waking up the user at the time corresponding to the second endpoint of the elastic wake-up interval, wherein the time corresponding to the first endpoint of the elastic wake-up interval is earlier than the time corresponding to the second endpoint.
According to the embodiment, through monitoring of the sleep index in the sleeping process of the user, the change of the sleep state of the user is judged and predicted by using the deep learning model, a matched sleep regulation scheme is automatically generated for the user in combination with the preset wake-up time set by the user, and the sleep state of the user is actively regulated and controlled based on the sleep regulation scheme to be matched with the sleep regulation scheme, so that the user is ensured to be in a shallow sleep state at the preset wake-up time, and the sleeping quality of the user is improved under the condition that the sleeping duration of the user is not influenced.
A third embodiment of the present disclosure provides a storage medium storing a computer program that, when executed by a processor, implements the steps of the deep learning model-based sleep regulation method provided by the first embodiment of the present disclosure.
A fourth embodiment of the present disclosure provides an apparatus, which at least includes a memory, and a processor, where the memory stores a computer program that when executed on the memory implements the steps of the deep learning model-based sleep regulation method provided by the first embodiment of the present disclosure.
In addition, the device of the embodiment may further include a device for performing sleep index collection and a device for performing sensory stimulation, or the device of the embodiment may communicate with or control the device for performing sleep index collection and the device for performing sensory stimulation.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present disclosure, and not for limiting the same; although the present disclosure has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present disclosure.

Claims (10)

1.一种基于深度学习的睡眠调控方法，其特征在于，包括：1. A sleep regulation method based on deep learning, characterized by comprising: 采集用户的当前睡眠指标；Collect the user's current sleep indicators; 将所述当前睡眠指标输入至深度学习模型，并且将所述当前睡眠指标存储至所述用户的个人信息数据库；Inputting the current sleep index into a deep learning model, and storing the current sleep index into a personal information database of the user; 所述深度学习模型结合所述个人信息数据库和所述用户的入睡时间，以所述当前睡眠指标作为输入，输出所述用户的睡眠状态预测曲线；The deep learning model combines the personal information database and the user's sleeping time, takes the current sleep index as input, and outputs a sleep state prediction curve for the user; 根据所述睡眠状态预测曲线和所述用户的预定唤醒时间，生成匹配所述用户的睡眠调控方案；generating a sleep regulation scheme matching the user according to the sleep state prediction curve and the scheduled wake-up time of the user; 根据所述睡眠调控方案对所述用户的睡眠状态进行基于预设方式的调控，使所述用户的睡眠状态与所述睡眠调控方案匹配。The sleep state of the user is regulated based on a preset manner according to the sleep regulation scheme, so that the sleep state of the user matches the sleep regulation scheme. 2.根据权利要求1所述的睡眠调控方法，其特征在于，所述深度学习模型结合所述个人信息数据库和所述用户的入睡时间，以所述当前睡眠指标作为输入，输出所述用户的睡眠状态预测曲线，包括：2. The sleep regulation method according to claim 1, characterized in that the deep learning model combines the personal information database and the user's sleep time, takes the current sleep index as input, and outputs the user's sleep state prediction curve, including: 所述深度学习模型根据所述当前睡眠指标预测下一时刻的睡眠指标，并将所述下一时刻的睡眠指标作为新的当前睡眠指标，重复进行下一时刻的睡眠指标的预测；The deep learning model predicts the sleep index at the next moment according to the current sleep index, and uses the sleep index at the next moment as the new current sleep index, and repeats the prediction of the sleep index at the next moment; 根据所述个人信息数据库和所述深度学习模型预测的所有下一时刻的睡眠指标，输出所述用户的睡眠状态预测曲线。According to the personal information database and all sleep indicators at the next moment predicted by the deep learning model, a sleep state prediction curve of the user is output. 3.根据权利要求1所述的睡眠调控方法，其特征在于，所述根据所述睡眠状态预测曲线和所述用户的预定唤醒时间，生成匹配所述用户的睡眠调控方案，包括：3. The sleep regulation method according to claim 1, characterized in that generating a sleep regulation scheme matching the user according to the sleep state prediction curve and the user's scheduled wake-up time comprises: 检测在所述睡眠状态预测曲线中所述预定唤醒时间所对应的睡眠状态是否为浅睡眠状态；Detecting whether the sleep state corresponding to the predetermined wake-up time in the sleep state prediction curve is a light sleep state; 在所述预定唤醒时间所对应的睡眠状态为浅睡眠状态的情况下，所述睡眠调控方案为空；When the sleep state corresponding to the scheduled wake-up time is a light sleep state, the sleep regulation scheme is empty; 在所述预定唤醒时间所对应的睡眠状态不为浅睡眠状态的情况下，在所述睡眠状态预测曲线的时间轴上确定距离所述预定唤醒时间最近的浅睡眠状态区域；When the sleep state corresponding to the scheduled wake-up time is not a light sleep state, determining a light sleep state area closest to the scheduled wake-up time on the time axis of the sleep state prediction curve; 根据所述浅睡眠状态区域与所述预定唤醒时间之间的先后顺序，确定所述睡眠调控方案为对所述用户的睡眠周期进行加速或减速。According to the sequence between the light sleep state area and the scheduled wake-up time, the sleep regulation scheme is determined to accelerate or decelerate the sleep cycle of the user. 4.根据权利要求1所述的睡眠调控方法，其特征在于，在所述根据所述睡眠调控方案对所述用户的睡眠状态进行基于预设方式的调控，使所述用户的睡眠状态与所述睡眠调控方案匹配之后，还包括：4. The sleep regulation method according to claim 1, characterized in that after regulating the sleep state of the user based on a preset manner according to the sleep regulation scheme so that the sleep state of the user matches the sleep regulation scheme, it further comprises: 实时采集当前睡眠指标；Collect current sleep indicators in real time; 根据所述当前睡眠指标确定用户当前的实际睡眠状态，根据所述实际睡眠状态对所述深度学习模型进行更新，并将所述当前睡眠指标存入所述个人信息数据库；Determine the user's current actual sleep state according to the current sleep index, update the deep learning model according to the actual sleep state, and store the current sleep index in the personal information database; 根据更新后的所述深度学习模型重新预测所述睡眠状态预测曲线；Re-predicting the sleep state prediction curve according to the updated deep learning model; 根据重新预测的所述睡眠状态预测曲线和所述用户的预定唤醒时间，重新生成睡眠调控方案，并基于所述重新生成的睡眠调控方案对所述用户的睡眠状态进行基于预设方式的调控。A sleep regulation scheme is regenerated according to the re-predicted sleep state prediction curve and the scheduled wake-up time of the user, and the sleep state of the user is regulated in a preset manner based on the regenerated sleep regulation scheme. 5.根据权利要求1所述的睡眠调控方法，其特征在于，还包括：5. The sleep regulation method according to claim 1, further comprising: 检测是否出现睡眠中断；Detect whether sleep interruptions occur; 在出现睡眠中断的情况下，重新确定所述用户的入睡时间；In the event of sleep interruption, re-determining the user's sleep time; 所述深度学习模型结合所述个人信息数据库和重新确定的所述用户的入睡时间，重新预测所述睡眠状态预测曲线；The deep learning model combines the personal information database and the re-determined sleeping time of the user to re-predict the sleep state prediction curve; 根据重新预测的所述睡眠状态预测曲线和所述用户的预定唤醒时间，重新生成睡眠调控方案，并基于所述重新生成的睡眠调控方案对所述用户的睡眠状态进行基于预设方式的调控。A sleep regulation scheme is regenerated according to the re-predicted sleep state prediction curve and the scheduled wake-up time of the user, and the sleep state of the user is regulated in a preset manner based on the regenerated sleep regulation scheme. 6.根据权利要求1所述的睡眠调控方法，其特征在于，所述根据所述睡眠状态预测曲线和所述用户的预定唤醒时间，生成匹配所述用户的睡眠调控方案，包括：6. The sleep regulation method according to claim 1, characterized in that generating a sleep regulation scheme matching the user according to the sleep state prediction curve and the user's scheduled wake-up time comprises: 获取用户的梦境偏好参数k以及睡眠调控极限时长N，其中，k∈[-1,1]；Obtain the user's dream preference parameter k and the sleep regulation limit duration N, where k∈[-1,1]; 确定睡眠调控额度M，其中，M＝|kN|；Determine the sleep regulation quota M, where M = |kN|; 根据所述梦境偏好参数k、所述预定唤醒时间以及所述睡眠状态预测曲线，确定符合所述用户的梦境偏好的第一睡眠周期；Determining a first sleep cycle that meets the dream preference of the user according to the dream preference parameter k, the scheduled wake-up time, and the sleep state prediction curve; 检测所述第一睡眠周期与所述用户未调控的睡眠周期之间的差值是否小于或等于所述睡眠调控额度；Detecting whether a difference between the first sleep cycle and the sleep cycle of the user that has not been adjusted is less than or equal to the sleep adjustment amount; 在所述第一睡眠周期与所述用户未调控的睡眠周期之间的差值小于或等于所述睡眠调控额度的情况下，所述睡眠调控方案为对所述用户的睡眠周期进行加速或减速，使所述睡眠周期经过调控后与所述第一睡眠周期重合；When the difference between the first sleep cycle and the unregulated sleep cycle of the user is less than or equal to the sleep regulation amount, the sleep regulation scheme is to accelerate or decelerate the sleep cycle of the user so that the regulated sleep cycle coincides with the first sleep cycle; 在所述第一睡眠周期与所述用户未调控的睡眠周期之间的差值大于所述睡眠调控额度的情况下，在所述睡眠状态预测曲线的时间轴上确定距离所述预定唤醒时间最近的浅睡眠状态区域，并根据所述浅睡眠状态区域与所述预定唤醒时间之间的先后顺序，确定所述睡眠调控方案为对所述用户的睡眠周期进行加速或减速。When the difference between the first sleep cycle and the user's unregulated sleep cycle is greater than the sleep regulation amount, a light sleep state area closest to the scheduled wake-up time is determined on the time axis of the sleep state prediction curve, and according to the order between the light sleep state area and the scheduled wake-up time, the sleep regulation plan is determined to accelerate or decelerate the user's sleep cycle. 7.根据权利要求1至6中任一项所述的睡眠调控方法，其特征在于，在无法调控所述用户的睡眠状态的情况下，还包括：7. The sleep regulation method according to any one of claims 1 to 6, characterized in that when the sleep state of the user cannot be regulated, it further comprises: 获取所述用户设置的弹性唤醒区间，所述预定唤醒时间位于所述弹性唤醒区间内；Acquire the flexible wake-up interval set by the user, wherein the scheduled wake-up time is within the flexible wake-up interval; 在所述弹性唤醒区间内确定所述用户的睡眠周期是否包括浅睡眠状态；Determining whether the sleep cycle of the user includes a light sleep state within the elastic wake-up interval; 在用户的睡眠周期包括浅睡眠状态的情况下，在所述浅睡眠状态对应的时间段内唤醒所述用户；In the case where the sleep cycle of the user includes a light sleep state, waking up the user within a time period corresponding to the light sleep state; 在用户的睡眠周期不包括浅睡眠状态的情况下，在所述预定唤醒时间所在的睡眠周期的上一个睡眠周期中用户进入深睡眠状态之前对所述用户进行基于预设方式的调控，以阻止用户进入深睡眠状态，或者，在所述弹性唤醒区间的第二端点对应的时间唤醒所述用户，其中，所述弹性唤醒区间的第一端点对应的时间早于所述第二端点对应的时间。In a case where the user's sleep cycle does not include a light sleep state, the user is regulated based on a preset method before the user enters a deep sleep state in a sleep cycle previous to the sleep cycle where the scheduled wake-up time is located, so as to prevent the user from entering the deep sleep state, or the user is woken up at a time corresponding to a second endpoint of the elastic wake-up interval, wherein the time corresponding to the first endpoint of the elastic wake-up interval is earlier than the time corresponding to the second endpoint. 8.一种基于深度学习的睡眠调控系统，其特征在于，包括：8. A sleep control system based on deep learning, characterized by comprising: 睡眠监测系统、运算存储系统以及感官刺激系统；其中，Sleep monitoring system, computing storage system and sensory stimulation system; among them, 所述睡眠监测系统用于采集用户的当前睡眠指标；The sleep monitoring system is used to collect the user's current sleep index; 所述运算存储系统用于将所述当前睡眠指标输入至深度学习模型，并且将所述当前睡眠指标存储至所述用户的个人信息数据库；所述深度学习模型结合所述个人信息数据库和所述用户的入睡时间，以所述当前睡眠指标作为输入，输出所述用户的睡眠状态预测曲线；根据所述睡眠状态预测曲线和所述用户的预定唤醒时间，生成匹配所述用户的睡眠调控方案；The computing storage system is used to input the current sleep index into the deep learning model, and store the current sleep index into the personal information database of the user; the deep learning model combines the personal information database and the user's sleep time, takes the current sleep index as input, and outputs the user's sleep state prediction curve; generates a sleep regulation plan that matches the user according to the sleep state prediction curve and the user's scheduled wake-up time; 所述感官刺激系统用于根据所述睡眠调控方案对所述用户的睡眠状态进行基于预设方式的调控，使所述用户的睡眠状态与所述睡眠调控方案匹配。The sensory stimulation system is used to regulate the user's sleep state based on a preset manner according to the sleep regulation scheme, so that the user's sleep state matches the sleep regulation scheme. 9.一种存储介质，存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现权利要求1至7中任一项所述的基于深度学习的睡眠调控方法的步骤。9. A storage medium storing a computer program, characterized in that when the computer program is executed by a processor, the steps of the deep learning-based sleep regulation method described in any one of claims 1 to 7 are implemented. 10.一种设备，至少包括存储器、处理器，所述存储器上存储有计算机程序，其特征在于，所述处理器在执行所述存储器上的计算机程序时实现权利要求1至7中任一项所述的基于深度学习的睡眠调控方法的步骤。10. A device comprising at least a memory and a processor, wherein a computer program is stored in the memory, wherein the processor implements the steps of the deep learning-based sleep regulation method described in any one of claims 1 to 7 when executing the computer program in the memory.
CN202311250571.6A 2023-09-26 2023-09-26 Deep learning-based sleep regulation method, system, storage medium and equipment Pending CN118398169A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311250571.6A CN118398169A (en) 2023-09-26 2023-09-26 Deep learning-based sleep regulation method, system, storage medium and equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311250571.6A CN118398169A (en) 2023-09-26 2023-09-26 Deep learning-based sleep regulation method, system, storage medium and equipment

Publications (1)

Publication Number Publication Date
CN118398169A true CN118398169A (en) 2024-07-26

Family

ID=91994391

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311250571.6A Pending CN118398169A (en) 2023-09-26 2023-09-26 Deep learning-based sleep regulation method, system, storage medium and equipment

Country Status (1)

Country Link
CN (1) CN118398169A (en)

Similar Documents

Publication Publication Date Title
US12315615B2 (en) Optimal sleep phase selection system
US7608041B2 (en) Monitoring and control of sleep cycles
US20190231256A1 (en) Apparatus and associated methods for adjusting a user's sleep
US20150348390A1 (en) Wake-up device
EP3517024B1 (en) An apparatus and associated methods for adjusting a group of users' sleep
US11395902B2 (en) System and method for waking a user up with a stimulus of varying intensity
CN112368776A (en) Controlling device output based on determined user condition
US11918525B1 (en) Sleep quality optimization using a controlled sleep surface
CN113639449A (en) Method and device for air conditioner control, air conditioner
CN111012132A (en) Sleep state adjusting method and device based on pillow and intelligent pillow
CN114431837B (en) Sleep state control method and device, sleep-assisting equipment and storage medium
CN118398169A (en) Deep learning-based sleep regulation method, system, storage medium and equipment
CN113094018A (en) Sleep control method and device, system, electronic equipment and storage medium thereof
JP2024012302A (en) Environmental control system and method
CN117839032A (en) Sleep monitoring method and device and intelligent bed
EP4664476A1 (en) Sleep management during medical examinations
US20250281103A1 (en) Apparatus and methods for active-feedback
CN120360511B (en) A sleep aid method and system
US20250054601A1 (en) A sleep tracking system
JP2023148754A (en) Equipment control system, equipment control method, and program
EP3932296A1 (en) System for sleep restriction therapy
CN119033346A (en) Sleep-aiding control method, system and equipment
WO2025095872A1 (en) A system and method developed for embedded devices and mobile platforms to facilitate falling sleep and improve the waking process
CN116869325A (en) Control method and system for temperature control mattress based on sleep monitoring
CN119701157A (en) Sleep light adjustment method, system, device and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination