CN110710962A - Sleep state detection method and device - Google Patents

Sleep state detection method and device Download PDF

Info

Publication number
CN110710962A
CN110710962A CN201911089684.6A CN201911089684A CN110710962A CN 110710962 A CN110710962 A CN 110710962A CN 201911089684 A CN201911089684 A CN 201911089684A CN 110710962 A CN110710962 A CN 110710962A
Authority
CN
China
Prior art keywords
sleep
heart rate
acceleration
characteristic
data
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
CN201911089684.6A
Other languages
Chinese (zh)
Inventor
张庆学
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.)
Beijing Calorie Information Technology Co Ltd
Original Assignee
Beijing Calorie Information Technology Co Ltd
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 Beijing Calorie Information Technology Co Ltd filed Critical Beijing Calorie Information Technology Co Ltd
Priority to CN201911089684.6A priority Critical patent/CN110710962A/en
Publication of CN110710962A publication Critical patent/CN110710962A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
    • A61B5/0205—Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
    • A61B5/024—Detecting, measuring or recording pulse rate or heart rate
    • A61B5/02438—Detecting, measuring or recording pulse rate or heart rate with portable devices, e.g. worn by the patient
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
    • A61B5/1118—Determining activity level
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/48—Other medical applications
    • A61B5/4806—Sleep evaluation
    • A61B5/4812—Detecting sleep stages or cycles
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/48—Other medical applications
    • A61B5/4806—Sleep evaluation
    • A61B5/4815—Sleep quality
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
    • A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
    • A61B5/6802—Sensor mounted on worn items
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
    • A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
    • A61B5/6802—Sensor mounted on worn items
    • A61B5/681—Wristwatch-type devices
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
    • A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
    • A61B5/683—Means for maintaining contact with the body
    • A61B5/6831—Straps, bands or harnesses
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
    • A61B5/6887—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient mounted on external non-worn devices, e.g. non-medical devices
    • A61B5/6898—Portable consumer electronic devices, e.g. music players, telephones, tablet computers
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
    • A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Veterinary Medicine (AREA)
  • Animal Behavior & Ethology (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Pathology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Biophysics (AREA)
  • Physiology (AREA)
  • Cardiology (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Physics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Fuzzy Systems (AREA)
  • Psychiatry (AREA)
  • Signal Processing (AREA)
  • Dentistry (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Multimedia (AREA)
  • Pulmonology (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

The invention discloses a sleep state detection method and a sleep state detection device. Wherein, the method comprises the following steps: acquiring heart rate data and acceleration data; determining heart rate characteristics according to the heart rate data, and determining acceleration characteristics according to the acceleration data; detecting sleep states according to the heart rate characteristics and the acceleration characteristics, wherein the sleep states comprise a sleep start, a sleep end, a deep sleep stage and a shallow sleep stage. The invention solves the technical problems that sleep monitoring in the related technology has certain limitation and does not support sleep detection all day long.

Description

Sleep state detection method and device
Technical Field
The invention relates to the field of sleep detection, in particular to a sleep state detection method and a sleep state detection device.
Background
At present, wearable products such as bracelet and intelligent wrist-watch are equipped with abundant sensor mostly, and wherein, most of them include six sensors and rhythm of the heart sensor, use six sensors to detect sleep state comparatively accurately, but, the consumption of gyroscope is higher relatively, to bracelet and the intelligent wrist-watch of all day operation, has undoubtedly provided certain challenge to duration. In addition, most of sleep detection methods based on the triaxial acceleration sensor can only detect night sleep and do not support afternoon nap detection, and the detection of all-day sleep has certain limitation.
In view of the above problems, no effective solution has been proposed.
Disclosure of Invention
The embodiment of the invention provides a sleep state detection method and a sleep state detection device, which at least solve the technical problems that sleep monitoring in the related technology has certain limitation and does not support all-day sleep detection.
According to an aspect of an embodiment of the present invention, there is provided a sleep state detection method, including: acquiring heart rate data and acceleration data; determining heart rate characteristics according to the heart rate data, and determining acceleration characteristics according to the acceleration data; detecting sleep states according to the heart rate characteristics and the acceleration characteristics, wherein the sleep states comprise a sleep start stage, a sleep end stage, a deep sleep stage and a shallow sleep stage.
Optionally, the acquiring the heart rate data and the acceleration data includes: detecting whether the wearable equipment is in a wearing state; and under the condition that the wearable equipment is in a wearing state, acquiring heart rate data and acceleration data recorded by the wearable equipment.
Optionally, determining the heart rate characteristic according to the heart rate data, and determining the acceleration characteristic according to the acceleration data includes: sampling the heart rate data, determining first sampling data, and taking the first sampling data as the heart rate characteristic; and/or sampling the acceleration data, determining second sampling data, and taking the second sampling data as the acceleration characteristic.
Optionally, determining the heart rate characteristic according to the heart rate data, and determining the acceleration characteristic according to the acceleration data includes: performing data extraction on the heart rate data according to an extraction window with preset duration; converting the heart rate data within the extraction window into a finite number of first discrete features, wherein the first discrete features include at least one of: heart rate change rate; taking the first discrete feature as the heart rate feature; and/or performing data extraction on the acceleration data according to an extraction window with preset duration; converting the acceleration data within the extraction window into a finite number of second discrete features, wherein the second discrete features include at least one of: activity intensity, activity duration, inactivity duration, activity and inactivity switching times; and taking the second discrete characteristic as the acceleration characteristic.
Optionally, determining the heart rate characteristic according to the heart rate data, and determining the acceleration characteristic according to the acceleration data includes: performing data extraction on the heart rate data according to an extraction window with preset duration; converting the difference and variance of the heart rate data extracted from a plurality of continuous extraction windows into a first conversion characteristic, wherein the first conversion characteristic comprises heart rate change rate, and the extraction windows extract the heart rate data at a preset frequency; taking the first conversion characteristic as the heart rate characteristic; and/or performing data extraction on the acceleration data according to an extraction window with preset duration; converting the difference and variance of the acceleration data extracted from a plurality of continuous extraction windows into second conversion characteristics, wherein the second conversion characteristics comprise an active state, an active time length, a static state, a static time length and the number of active and static switching times, and the extraction windows extract the acceleration data at a preset frequency; and taking the second conversion characteristic as the acceleration characteristic.
Optionally, the sleep state further includes: the sleep stages, secondary sleep, in which sleep is in, begin; detecting a sleep state according to the heart rate characteristic and the acceleration characteristic, comprising: judging whether the current sleep state is in, carrying out staged detection on a plurality of sleep stages of the sleep state under the condition of the sleep state to determine the sleep stage, and then carrying out detection of sleep completion; under the condition that the sleep state is not achieved, whether the time interval from the last sleep exceeds a preset interval threshold value or not is judged, and under the condition that the time interval exceeds the preset interval threshold value, the sleep start is detected; and under the condition that the time interval does not exceed the preset interval threshold, detecting the start of the secondary sleep.
Optionally, detecting a sleep state according to the heart rate characteristic and the acceleration characteristic includes: detecting whether the sleep state is sleep start or not according to the heart rate characteristics and the acceleration characteristics; detecting sleep onset of the sleep state based on the heart rate characteristic and the acceleration characteristic comprises: inputting the heart rate characteristic or the acceleration characteristic into a first logic condition group, and outputting a detection result by the first logic condition group, wherein the first logic condition group comprises a plurality of first logic conditions, and wherein the first logic condition group comprises at least one of the following conditions: whether the static duration reaches a first preset static duration or not, whether the activity duration is lower than a first preset activity duration or not, whether the activity amplitude is smaller than a first preset amplitude or not and whether the activity intensity is lower than a first preset activity intensity or not; or, detecting whether the sleep state is sleep onset according to the heart rate characteristic and the acceleration characteristic includes: inputting the heart rate characteristics or the acceleration characteristics into a first recognition model, and outputting a first detection result by the first recognition model, wherein the first recognition model is formed by training a plurality of groups of training data, and each group of training data comprises the heart rate characteristics or the acceleration characteristics and a corresponding first detection result.
Optionally, detecting a plurality of sleep stages in which a sleep state is located according to the heart rate characteristic and the acceleration characteristic, further includes: detecting a plurality of sleep stages in a sleep state in stages according to the heart rate characteristics and the acceleration characteristics, wherein the sleep stages include at least one of: a light sleep stage, a deep sleep stage; staging detection of multiple sleep stages in a sleep state according to the heart rate signature and the acceleration signature comprises: and judging whether the heart rate characteristic and the acceleration characteristic are in a preset threshold range or not, and determining the sleep stage to be the sleep stage corresponding to the preset threshold range under the condition of being in the preset threshold range.
Optionally, before performing the staged detection on the multiple sleep stages in the sleep state according to the heart rate characteristic and the acceleration characteristic, the method further includes: predicting the preset threshold value ranges corresponding to different sleep stages in the current sleep state according to the historical record of the last sleep state, wherein the historical record comprises heart rate characteristics and acceleration characteristics of the user in a plurality of different sleep stages; or, under the condition of initialization, clearing a history record, and predicting the preset threshold value ranges corresponding to different sleep stages in the current sleep state according to a priori values given by the initialization, wherein the history record comprises heart rate characteristics and acceleration characteristics of the user in a plurality of different sleep stages.
Optionally, detecting a sleep state according to the heart rate characteristic and the acceleration characteristic, further comprising: detecting the onset of a second sleep according to the heart rate characteristic and the acceleration characteristic; detecting onset of a secondary sleep based on the heart rate characteristic and the acceleration characteristic comprises: inputting the heart rate characteristic or the acceleration characteristic into a second logic condition group, and outputting a detection result by the second logic condition group, wherein the second logic condition group comprises a plurality of second logic conditions; the second logic condition group comprises whether the static time length reaches a second preset static time length or not, whether the activity time length is lower than a second preset activity time length or not, whether the activity amplitude is smaller than a second preset amplitude or not and whether the activity intensity is lower than a second preset activity intensity or not; or, detecting onset of a secondary sleep according to the heart rate characteristic and the acceleration characteristic includes: and inputting the heart rate characteristics or the acceleration characteristics into a second recognition model, and outputting a second detection result by the second recognition model, wherein the second recognition model is formed by training a plurality of groups of training data, and each group of training data comprises the heart rate characteristics or the acceleration characteristics of the second sleep and the corresponding second detection result.
Optionally, before detecting the start of the second sleep according to the heart rate characteristic and the acceleration characteristic, the heart rate characteristic and the acceleration characteristic in the sleep state before the waking period corresponding to the second sleep are deleted.
Optionally, detecting a sleep state according to the heart rate characteristic and the acceleration characteristic includes: detecting whether the sleep state is sleep termination according to the heart rate characteristics and the acceleration characteristics; detecting the end of sleep in the sleep state based on the heart rate characteristic and the acceleration characteristic comprises: inputting the heart rate characteristic or the acceleration characteristic into a third logic condition set, and outputting a detection result by the third logic condition set, wherein the third logic condition set comprises a plurality of third logic conditions, and wherein the third logic condition set comprises at least one of the following conditions: whether the activity duration reaches a third preset activity duration, the activity amplitude reaches a third preset amplitude, and whether the activity intensity reaches a third preset activity intensity; or, detecting whether the sleep state is sleep end according to the heart rate characteristic and the acceleration characteristic includes: and inputting the heart rate characteristics or the acceleration characteristics into a third recognition model, and outputting a third detection result by the third recognition model, wherein the third recognition model is formed by training a plurality of groups of training data, and each group of training data comprises the heart rate characteristics or the acceleration characteristics and a corresponding third detection result.
According to another aspect of the embodiments of the present invention, there is also provided a sleep state detection apparatus, including: the acquisition module is used for acquiring heart rate data and acceleration data; the determining module is used for determining heart rate characteristics according to the heart rate data and determining acceleration characteristics according to the acceleration data; the detection module is used for detecting sleep states according to the heart rate characteristics and the acceleration characteristics, wherein the sleep states comprise a sleep start stage, a sleep end stage, a deep sleep stage and a shallow sleep stage.
According to another aspect of the embodiments of the present invention, there is also provided a storage medium, where the storage medium includes a stored program, and when the program runs, the apparatus where the storage medium is located is controlled to execute any one of the foregoing sleep state detection methods.
According to another aspect of the embodiments of the present invention, there is also provided a processor, configured to execute a program, where the program executes to perform the sleep state detection method described in any one of the above.
In the embodiment of the invention, heart rate data and acceleration data are acquired; determining heart rate characteristics according to the heart rate data, and determining acceleration characteristics according to the acceleration data; the sleep state is detected according to the heart rate characteristics and the acceleration characteristics, wherein the sleep state comprises a sleep start mode, a sleep end mode, a deep sleep stage mode and a shallow sleep stage mode, the heart rate characteristics and the acceleration characteristics are determined according to the heart rate data and the acceleration data, the sleep state is detected according to the heart rate characteristics and the acceleration characteristics, the purpose of detecting the sleep state at any time is achieved, the technical effect of all-day sleep detection is achieved, and the technical problems that sleep monitoring in the related art has certain limitation and all-day sleep detection is not supported are solved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:
FIG. 1 is a flow chart of a sleep state detection method according to an embodiment of the present invention;
FIG. 2 is a flow chart of a sleep state detection method according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a sleep state detection apparatus according to an embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
In accordance with an embodiment of the present invention, there is provided a method embodiment of a sleep state detection method, it should be noted that the steps illustrated in the flowchart of the accompanying drawings may be performed in a computer system such as a set of computer executable instructions, and that while a logical order is illustrated in the flowchart, in some cases the steps illustrated or described may be performed in an order different than here.
Fig. 1 is a flowchart of a sleep state detection method according to an embodiment of the present invention, as shown in fig. 1, the method includes the following steps:
step S102, heart rate data and acceleration data are obtained;
step S104, determining heart rate characteristics according to the heart rate data, and determining acceleration characteristics according to the acceleration data;
and S106, detecting sleep states according to the heart rate characteristics and the acceleration characteristics, wherein the sleep states comprise a sleep start stage, a sleep end stage, a deep sleep stage and a shallow sleep stage.
Through the steps, the acquisition of heart rate data and acceleration data can be realized; determining heart rate characteristics according to the heart rate data, and determining acceleration characteristics according to the acceleration data; the sleep state is detected according to the heart rate characteristics and the acceleration characteristics, wherein the sleep state comprises a sleep start mode, a sleep end mode, a deep sleep stage mode and a shallow sleep stage mode, the heart rate characteristics and the acceleration characteristics are determined according to the heart rate data and the acceleration data, the sleep state is detected according to the heart rate characteristics and the acceleration characteristics, the purpose of detecting the sleep state at any time is achieved, the technical effect of all-day sleep detection is achieved, and the technical problems that sleep monitoring in the related art has certain limitation and all-day sleep detection is not supported are solved.
In the related art, whether the sleep state is in can be directly judged through the heart rate data and the acceleration data, but because the identification of the heart rate data and the acceleration data has certain errors, the problem that the sleep state is judged directly through the heart rate data and the acceleration data is inaccurate is caused. According to the method and the device, the sleep state is determined according to the characteristics of the heart rate data and the acceleration data, such as the change rate, the state duration and the like, so that the accuracy rate of data acquisition is prevented from influencing the accuracy rate and the accuracy rate of sleep state determination. And the sleep state of the user cannot be judged in the afternoon nap state or the sleep state under an unspecified condition by directly judging whether the user is in the sleep state or not under the specified condition, such as the night time period, through the heart rate data and the acceleration data, so that the sleep quality of the user cannot be comprehensively determined, and the requirements of the user are met.
The aforesaid acquires rhythm of heart data and acceleration data, can be the user's that acquires wearing equipment collection rhythm of heart data and acceleration data through wearing equipment, and above-mentioned wearing equipment can be intelligent bracelet, smart watch, smart mobile phone etc. and above-mentioned rhythm of heart data is wearing equipment affiliated user's rhythm of the heart data, can demonstrate user's healthy state to a certain extent through rhythm of the heart data. Above-mentioned acceleration data are the acceleration data of the position of wearing of smart machine on above-mentioned user, and the position of wearing of different wearing equipment is different, for example, smart watch and smart bracelet can wear usually in the wrist position, and above-mentioned smart mobile phone need tie up in the first half section or the back half section of arm with the help of the cell-phone bandage when wearing, and above-mentioned smart mobile phone can also tie up on thigh or shank through the cell-phone bandage. Different body positions are different in accelerated speed under different sleep states, so that before the sleep state is detected according to the heart rate characteristics and the accelerated speed characteristics, the accelerated speed data standard of the body position worn by the wearable device can be acquired, and the sleep state is effectively detected through the accelerated speed data characteristics.
The heart rate characteristics determined according to the heart rate data may be determined according to the variation condition of the heart rate data in each heart rate interval. The heart rate characteristic may be an active state, a hyper-active state, a hypo-active state, a resting state, etc. Specifically, the user is determined to be in the active state in a high heart rate interval in which the heart rate data is continuously kept in the active state, and the user is determined to be in the super-strong active state in a super-strong heart rate interval in which the heart rate data is continuously kept in the super-strong active state. Since the interaction state of the normal person in the sleep state can be a still state or a low intensity state, it can be determined whether the user is in the sleep state by determining whether the user is in the still state or the low intensity state. In another aspect, whether the user enters the sleep state or not can be determined according to the change rate of the heart rate data, the change rate of the heart rate does not exceed a preset heart rate threshold value when a normal person is in the sleep state, the user is considered not to be in the sleep state under the condition that the measured change rate of the heart rate of the user exceeds the preset heart rate threshold value, and the user is considered to be in the sleep state under the condition that the measured change rate of the heart rate of the user does not exceed the preset heart rate threshold value.
The acceleration data determines acceleration characteristics, and the amplitude and the speed of the user action can be determined according to the threshold range of the acceleration, so as to determine the activity state of the user, for example, the user is determined to be in the activity state within the numerical range of the activity state of the acceleration data, and similarly, whether the user is in the super-strong activity state, the low-intensity activity state and the still activity state can be determined. The activity state can be divided into more levels according to needs, and the more the activity state levels are, the more accurate the sleep state of the user is determined.
Detecting sleep states according to the heart rate characteristics and the acceleration characteristics, wherein the sleep states comprise a sleep start, a sleep end, a deep sleep stage and a shallow sleep stage. The sleep state detection at least includes when the sleep starts and when the sleep ends, that is, at least includes detecting the start of the sleep and detecting the end of the sleep. Various sleep stages in the sleep state are also included, including deep sleep stages, shallow sleep stages. When the sleep starts, the heart rate of the user is stabilized from a region with larger fluctuation, the motion of the user is that the acceleration data is from large to small until the user is still, after the user is still for a period of time, the change of the heart rate of the user is determined to be in accordance with the heart rate characteristic of falling asleep, the user is determined to fall asleep, the sleep state can be continuously monitored in the sleep process, as the sleep comprises a plurality of sleep stages, such as a deep sleep stage and a shallow sleep stage, the body states of the user are different according to different behaviors of the user in different sleep stages, and dreams, dream words, turning over and the like are possible in the deep sleep stage. In particular, heart rate characteristics and acceleration characteristics at different sleep stages require a large amount of experimental data support. The heart rate characteristic or the acceleration characteristic can be determined from a large amount of experimental data through a deep learning model or a machine learning model.
Specifically, the learning model can be trained by data features of a large amount of data and corresponding sleep states, so that whether the data features are in the corresponding sleep states or not can be determined according to the learning model. The learning model is formed by training a plurality of groups of training data, and each group of training data comprises heart rate characteristics or acceleration characteristics and a corresponding sleep state.
Optionally, the acquiring heart rate data and acceleration data includes: detecting whether the wearable equipment is in a wearing state; and under the condition that the wearable device is in a wearing state, acquiring heart rate data and acceleration data recorded by the wearable device.
Above-mentioned when obtaining heart rate data and acceleration data, to whether wearable equipment is in the wearing state, under wearable equipment is in the wearing state's the condition, just can obtain heart rate data and acceleration data through above-mentioned wearable equipment, above-mentioned wearable equipment must have the function and the device of obtaining user's heart rate data and reading data with higher speed of course. Above-mentioned wearable equipment can also be virtual reality VR equipment, augmented reality AR equipment, or other intelligent wearing equipment. The intelligent wearable equipment can be automatic temperature control clothes, automatic color changing clothes and the like, and can be worn on the body of a user.
Optionally, determining the heart rate characteristic according to the heart rate data, and determining the acceleration characteristic according to the acceleration data includes: sampling the heart rate data, determining first sampling data, and taking the first sampling data as heart rate characteristics; and/or sampling the acceleration data, determining second sampling data, and taking the second sampling data as the acceleration characteristic.
The heart rate characteristic is determined according to the heart rate data, when the acceleration characteristic is determined according to the acceleration data, the heart rate data or the acceleration data can be counted according to time and then sampled, and the down-sampled data is used as characteristic data. The sampling mode may be a preset time duration and sampling according to a certain frequency, or may be a sampling mode performed when the change condition of the detected data exceeds a certain threshold, or may be a variety of modes such as manual control sampling.
In the process of determining the heart rate data and the acceleration data according to the heart rate data and the acceleration data, data extraction can be carried out on the heart rate data according to an extraction window with preset duration; converting the heart rate data in the extraction window into a finite number of first discrete features, wherein the first discrete features include at least one of: heart rate change rate; taking the first discrete feature as a heart rate feature; and/or performing data extraction on the acceleration data according to an extraction window with preset duration; converting the acceleration data within the extraction window into a finite number of second discrete features, wherein the second discrete features include at least one of: activity intensity, activity duration, inactivity duration, activity and inactivity switching times; and taking the second discrete characteristic as an acceleration characteristic.
In the process of determining the heart rate data and the acceleration data according to the heart rate data and the acceleration data, data extraction can be carried out on the heart rate data according to an extraction window with preset duration; converting the difference and variance of the heart rate data extracted by a plurality of continuous extraction windows into a first conversion characteristic, wherein the first conversion characteristic comprises heart rate change rate, and the extraction windows extract the heart rate data at a preset frequency; taking the first conversion feature as a heart rate feature; and/or performing data extraction on the acceleration data according to an extraction window with preset duration; converting the difference and variance of the acceleration data extracted by a plurality of continuous extraction windows into second conversion characteristics, wherein the second conversion characteristics comprise an active state, an active time length, a static state, a static time length and the number of times of active and static switching, and the extraction windows extract the acceleration data at a preset frequency; and taking the second conversion characteristic as an acceleration characteristic.
Optionally, the sleep state may further include: the sleep stages, secondary sleep, in which sleep is in, begin; detecting a sleep state according to the heart rate characteristics and the acceleration characteristics, comprising: judging whether the current sleep state is in, carrying out staged detection on a plurality of sleep stages of the sleep state under the condition of the sleep state to determine the sleep stage, and then carrying out detection of sleep completion; under the condition that the sleep state is not achieved, whether the time interval from the last sleep exceeds a preset interval threshold value or not is judged, and under the condition that the time interval exceeds the preset interval threshold value, the sleep start is detected; and under the condition that the time interval does not exceed a preset interval threshold, detecting the start of the secondary sleep.
The sleep quality of the user in the sleep process can be determined by the sleep stages of the sleep so as to adapt to diversified user requirements. The onset of the second sleep is a determination of the brief waking activity of the user during sleep, e.g., at night. In the process of the secondary sleep, the heart rate characteristic and the acceleration characteristic of the sleep start are not obvious as those of the sleep start for the first time, so that the change standard of the heart rate characteristic and the acceleration characteristic can be properly reduced for the detection of the secondary sleep start, wherein the change standard of the heart rate characteristic and the acceleration characteristic can determine that the user starts to sleep and enters the sleep. The criterion to be higher is lowered by a little, for example, the change criterion is that the stationary state time is t1, and can be shortened to t2, t2< t 1. The heart rate characteristics and the acceleration characteristics of a plurality of sleep states in the sleep are different from the heart rate characteristics and the acceleration characteristics of the start of sleep and the end of sleep, and the corresponding heart rate characteristics and the corresponding speed characteristics of different sleep states in the sleep can be determined according to the mode of the learning model.
Optionally, detecting the sleep state according to the heart rate characteristic and the acceleration characteristic includes: detecting whether the sleep state is sleep start or not according to the heart rate characteristics and the acceleration characteristics; detecting sleep onset of the sleep state based on the heart rate characteristic and the acceleration characteristic includes: inputting the heart rate characteristic or the acceleration characteristic into a first logic condition group, and outputting a detection result by the first logic condition group, wherein the first logic condition group comprises a plurality of first logic conditions, and the first logic condition group comprises at least one of the following conditions: whether the static duration reaches a first preset static duration or not, whether the activity duration is lower than a first preset activity duration or not, whether the activity amplitude is smaller than a first preset amplitude or not and whether the activity intensity is lower than a first preset activity intensity or not; or, the detecting whether the sleep state is sleep onset according to the heart rate characteristic and the acceleration characteristic includes: the heart rate characteristics or the acceleration characteristics are input into a first recognition model, and a first detection result is output by the first recognition model, wherein the first recognition model is formed by training a plurality of groups of training data, and each group of training data comprises the heart rate characteristics or the acceleration characteristics and the corresponding first detection result.
The first logic condition set may be a plurality of first logic conditions, for example, determining whether the rest duration reaches a first preset rest duration, in the case of yes, it is determined whether the activity duration is lower than a first preset activity duration, in the case of yes, it is determined that the activity amplitude is smaller than a first preset amplitude, in the case of yes, determining whether the activity intensity is lower than a first preset activity intensity, and in the case of yes, determining to go to sleep, it is stated that all of the four first logic conditions described above must be met to determine that the user is entering a sleep state, of course, in other embodiments of this embodiment, the first logic condition may not be four, the first logical condition may not be any of the four, the four logical conditions may not be determined according to the logical relationship, for example, any three of the four first logic conditions described above may be satisfied to consider the user to go to sleep. Or otherwise, etc. The specific logic conditions, or the relationship of the logic conditions, may be set by the user personally, or determined empirically.
Optionally, detecting a plurality of sleep stages in which the sleep state is located according to the heart rate characteristic and the acceleration characteristic, further includes: detecting a plurality of sleep stages in the sleep state in stages according to the heart rate characteristics and the acceleration characteristics, wherein the sleep stages comprise at least one of the following: a light sleep stage, a deep sleep stage; the staged detection of multiple sleep stages in the sleep state based on the heart rate signature and the acceleration signature comprises: and judging whether the heart rate characteristic and the acceleration characteristic are in a preset threshold range, and determining the sleep stage to be the sleep stage corresponding to the preset threshold range under the condition that the heart rate characteristic and the acceleration characteristic are in the preset threshold range.
Before the above-mentioned a plurality of sleep stages in the sleep state are detected stage by stage according to the heart rate characteristic and the acceleration characteristic, still include: predicting preset threshold ranges corresponding to different sleep stages in the current sleep state according to a historical record of the last sleep state, wherein the historical record comprises heart rate characteristics and acceleration characteristics of a user in a plurality of different sleep stages; or, under the condition of initialization, clearing the historical record, and predicting the preset threshold value ranges corresponding to different sleep stages in the current sleep state according to the prior value given by the initialization, wherein the historical record comprises heart rate characteristics and acceleration characteristics of the user in a plurality of different sleep stages.
The preset threshold value of the heart rate characteristic is given by an empirical value/prior value for the first time, namely, initially given, and is obtained by subsequent learning according to a heart rate statistic value of the previous whole sleep, such as a mean value or a median value; the acceleration activity threshold is given by experience/prior value for the first time, namely initialization, and is obtained by learning according to the activity statistical result of the previous whole sleep, for example, a histogram statistical result and the like. The empirical/prior values may be obtained by any means including, but not limited to, manually collected data, empirical data, etc. And judging the deep sleep according to the heart rate and the activity threshold.
The detection of the onset of the secondary sleep is similar to the above-mentioned detectable onset of sleep, and specifically, as follows, optionally, the detection of the sleep state according to the heart rate characteristic and the acceleration characteristic further includes: detecting the beginning of the second sleep according to the heart rate characteristics and the acceleration characteristics; detecting onset of a secondary sleep based on the heart rate characteristic and the acceleration characteristic comprises: inputting the heart rate characteristic or the acceleration characteristic into a second logic condition group, and outputting a detection result by the second logic condition group, wherein the second logic condition group comprises a plurality of second logic conditions; the second logic condition group comprises whether the static duration reaches a second preset static duration or not, whether the activity duration is lower than a second preset activity duration or not, whether the activity amplitude is smaller than a second preset amplitude or not and whether the activity intensity is lower than a second preset activity intensity or not; or, detecting onset of a secondary sleep according to the heart rate characteristic and the acceleration characteristic includes: and inputting the heart rate characteristics or the acceleration characteristics into a second recognition model, and outputting a second detection result by the second recognition model, wherein the second recognition model is formed by training a plurality of groups of training data, and each group of training data comprises the heart rate characteristics or the acceleration characteristics of the second sleep and the corresponding second detection result.
Before the start of the second sleep is detected according to the heart rate characteristic and the acceleration characteristic, the heart rate characteristic and the acceleration characteristic in the sleep state before the waking period corresponding to the second sleep are deleted.
The influence of the heart rate characteristic and the acceleration characteristic of the previous sleep state on the judgment of the current sleep state is eliminated, so that the accuracy of the determination of the sleep state is improved.
Optionally, detecting the sleep state according to the heart rate characteristic and the acceleration characteristic includes: detecting whether the sleep state is the end of sleep according to the heart rate characteristic and the acceleration characteristic; detecting the end of sleep in the sleep state based on the heart rate characteristic and the acceleration characteristic includes: inputting the heart rate characteristic or the acceleration characteristic into a third logic condition group, and outputting a detection result by the third logic condition group, wherein the third logic condition group comprises a plurality of third logic conditions, and the third logic condition group comprises at least one of the following conditions: whether the activity duration reaches a third preset activity duration, the activity amplitude reaches a third preset amplitude, and whether the activity intensity reaches a third preset activity intensity; or, the detecting whether the sleep state is the end of sleep according to the heart rate characteristic and the acceleration characteristic includes: and inputting the heart rate characteristics or the acceleration characteristics into a third recognition model, and outputting a third detection result by the third recognition model, wherein the third recognition model is formed by training a plurality of groups of training data, and each group of training data comprises the heart rate characteristics or the acceleration characteristics and the corresponding third detection result.
It should be noted that this embodiment also provides an alternative implementation, which is described in detail below.
In the embodiment, the sleep state is detected by using a logic judgment or machine learning method by utilizing the relevant action characteristics of the triaxial acceleration and the change characteristics of the heart rate value, and the method can be applied to a bracelet or a smart watch provided with a heart rate sensor, a triaxial acceleration sensor or an integrated sensor comprising the triaxial acceleration sensor. The user only needs to wear bracelet or intelligent wrist-watch when sleep, and equipment can the current state of automated inspection user, sleep or awake etc to distinguish the degree of depth sleep and the shallow sleep state of locating in the sleep, thereby make the user need not to control the sleep condition that the wrist-watch also can accurately know oneself.
The existing sleep state detection method is partially based on six-axis sensors, including a three-axis acceleration sensor and a three-axis gyroscope or an electrocardio sensor, more detailed characteristics of hands and wrists can be acquired by using the gyroscope sensors, but the power consumption is high, and the endurance is influenced; the electrocardio sensor is not suitable for being worn for a long time and has higher cost. In addition, most of the identification methods based on the triaxial accelerometer only support night sleep detection, and the detection at night and the sleep support all day are less, so that the identification method has certain limitation on the practical application effect.
The arrangement thought of the embodiment is to collect data of the acceleration sensor and the heart rate sensor, process the data point by point, extract the change characteristics of the acceleration and the heart rate, select a proper logic judgment and machine learning model according to the current and historical heart rate value change conditions, input the characteristics and obtain a corresponding sleep state result. Data samples of more scenes can be enriched, and the application range and performance of the model are improved; and screening model characteristics to obtain a more compact model and the like.
The embodiment provides a human body sleep state detection method based on a bracelet or a smart watch, and particularly, a user wears the bracelet or the smart watch during sleep without any operation, and the bracelet or the smart watch can automatically detect whether the user starts sleeping or when the user starts sleeping according to the action state and the heart rate characteristics of the user; whether behaviors such as waking or rising night exist in the sleep period is detected, and if the behaviors such as waking or rising night are detected, the behaviors are marked as a waking state; and finally, detecting whether the user is awake according to the activity state and the heart rate change of the user, and recording the whole sleep state, the duration, the deep and shallow sleep stage and the like. In addition, the method also supports the detection of short-time sleep such as afternoon nap.
Fig. 2 is a flowchart of a sleep state detection method according to an embodiment of the present invention, as shown in fig. 2. The whole process starts from the point-by-point data collection of the acceleration sensor, and the data sequentially pass through all modules of the algorithm to finally obtain the current detection result.
The key technology of the embodiment lies in the processing algorithm of the data of the acceleration sensor and the heart rate sensor, and the algorithm is mainly divided into a module 1 (wearing detection), a module 2 (feature extraction), a module 3 (sleep start time detection), a module 4 (secondary sleep start time detection), a module 5 (sleep staging detection) and a module 6 (sleep end time detection) in fig. 2. The detailed description is as follows:
module 1: wear detection;
the detection method comprises the steps of detecting the wearing state of the bracelet or the smart watch, returning to the wearing or non-wearing state, judging whether the bracelet or the smart watch is worn or not by the aid of the change condition of the heart rate value and the acceleration data, and judging whether the bracelet or the smart watch is worn or not by the aid of long-time invariance of the acceleration data. Wearing of bracelet or smart watch and not wearing state and sleep state's detection direct correlation.
And (3) module 2: extracting characteristics;
according to the acceleration data in a window with a fixed length, relevant sleep features are extracted, the features can be divided into two types, one type is a direct feature, namely the acceleration data and the heart rate data in the window are directly used as the features, or appropriate down-sampling is carried out according to actual needs, the dimensionality of a feature vector is reduced, for example, the sampling rate of the original acceleration data is 50Hz, and the sampling rate of the acceleration data can be reduced to 5Hz or even lower through a sampling or re-sampling mode, so that the calculation complexity and the time cost are reduced; the other is indirect feature, that is, the acceleration and heart rate data in the window are converted into a finite number of discrete features by means of feature extraction, wherein the types of the discrete features include, but are not limited to, activity intensity, activity duration, rest duration, activity and rest switching times, heart rate change rate, and the like. Statistical methods for heart rate change rate characteristics include, but are not limited to, an upward/downward trend of the heart rate value within a fixed time period, a change interval length, a jump amplitude, and the like.
In a possible case, the acceleration data of a 10-second window is fixedly intercepted every time, the characteristics of the difference value of the maximum value and the minimum value of the acceleration in the window, the change difference value between adjacent windows, the window acceleration, the heart rate value variance and the like are counted, and the current window is determined to be in an active state or a static state by combining the discrete characteristics and logic judgment conditions.
In another possible case, the features such as difference values and variances extracted from a plurality of continuous windows are converted into conversion features such as active state, active time length, static state and static time length, active and static switching times, heart rate change rate and the like, and feature dimensions are further compressed.
And a module 3: detecting the sleep starting time;
and detecting the sleep starting time, namely obtaining discrete characteristics or conversion characteristics through a module 2 (characteristic extraction) and detecting the sleep starting time. The detection method mainly comprises two detection methods, wherein the first detection method adopts a traditional logic judgment method, uniformly judges the sleep starting time through multilayer logic conditions, such as maximum continuous static time, the number (number) of sections with static time exceeding a first threshold, the number (number) of sections with active time or amplitude lower than a second threshold, and the like, and simultaneously considers the time consumed for judging the sleep starting state for advancing the starting time so as to enable the detection result to be more accurate. Secondly, inputting the discrete features or the conversion features extracted by the module 2 (feature extraction) and the labeled sleep state into a training model by using a machine learning method, training to obtain a relation model 1 of the discrete features or the conversion features and the sleep state, and only inputting the discrete features or the conversion features of the current time or a window when the model is applied, wherein the output of the model 1 is the current sleep state, so that the sleep starting time is obtained, and the structure of the model 1 comprises but not limited to a decision tree model, a random forest model, a support vector machine model, a neural network model and the like.
The traditional logic judgment method and the machine learning method can be used simultaneously, the confidence coefficient evaluation is carried out on the output of the two methods, and the method corresponding to the result with higher confidence coefficient is adopted, so that the detection accuracy can be improved.
And (4) module: detecting the starting time of the second sleep;
mainly refers to the detection of the beginning time of the second sleep after getting up or waking. For the detection of the onset of a second sleep, the approach of block 3 (sleep onset detection) can also be used, with the only difference that, since the time of the night or wake is typically short, the extracted discrete features or transition features may contain sleep features before the night or wake, and therefore the sleep features before the night or wake need to be filtered and deleted at the time of the second sleep detection. In addition, considering the factors that it may be easier to fall asleep after getting up or awake, etc., the detection logic and threshold for the secondary sleep onset time may be relatively loose.
And a module 5: detecting sleep stages;
the sleep stage detection is to judge each stage of sleep according to the activity amount and the heart rate change condition in sleep, and the method mainly applied comprises the detection of the activity amount and the action amplitude in sleep, the heart rate change trend and the like. If the heart rate value is lower than the third threshold value and the activity amount in the period of time is lower than the fourth threshold value (namely, very small or almost no movement), the patient is in a deep sleep stage, otherwise, the patient is in a shallow sleep stage. The third threshold value can be determined by the average value or the median value of the whole sleep period, or can be drawn by using the prior value of the off-line data, and the fourth threshold value can be drawn by using the prior value of the off-line data, and then the third threshold value and the fourth threshold value are continuously learned and adjusted in the application. The prior value is obtained by a method including, but not limited to, manually collected data, empirical data, and the like. Meanwhile, the sleep stage is assisted and judged by using the sleep cycle regularity, for example, each night sleep process comprises a limited number of sleep cycles, and each cycle lasts for 90-110 min.
And a module 6: detecting the sleep ending time;
and detecting the end time of the sleep, namely obtaining discrete characteristics or conversion characteristics through a module 2 (characteristic extraction) and detecting the end time of the sleep. The detection method mainly comprises two detection methods, wherein the first detection method is characterized in that a traditional logic judgment method is used, the end time of sleep is judged uniformly through multilayer logic conditions, such as continuous activity time, the number (number) of sections with activity time exceeding a fifth threshold, the number (number) of sections with activity intensity exceeding a sixth threshold, and the like, and the time consumed by judging the sleep end state is also considered to be used for advancing the sleep end time, so that the detection result is more accurate. Secondly, inputting the discrete features or the conversion features obtained by the module 2 (feature extraction) and the labeled sleep state into a training model by using a machine learning method, training to obtain a relation model 2 of the discrete features or the conversion features and the sleep state, and only inputting the discrete features or the conversion features of the current time or the window when the method is applied, wherein the output of the model 2 is the current sleep state, so that the end time of the sleep is obtained. The structure of model 2 includes, but is not limited to, a decision tree model, a random forest model, a support vector machine model, a neural network model, and the like.
The traditional logic judgment method and the machine learning method can be used simultaneously, the confidence coefficient evaluation is carried out on the output of the two methods, and the method corresponding to the result with higher confidence coefficient is adopted, so that the detection accuracy can be improved.
This embodiment is through wearing bracelet or intelligent wrist-watch, and length, degree of depth sleep and light sleep etc. are long in the sleep state, sleep of utilizing acceleration sensor and rhythm of the heart sensor automatic identification user, in addition, still support the detection of getting up to night and afternoon nap, so can promote the performance of bracelet or wrist-watch to the experience of user to bracelet or wrist-watch is felt to a certain extent is promoted. Can promote the user and feel to the experience of bracelet or wrist-watch, can increase the time of endurance of bracelet wrist-watch to a certain extent simultaneously.
The embodiment can also be applied to other wearable devices, sports devices, motion sensing game devices and the like.
Fig. 3 is a schematic diagram of a sleep state detection apparatus according to an embodiment of the present invention, and as shown in fig. 3, according to another aspect of the embodiment of the present invention, there is also provided a sleep state detection apparatus including: an acquisition module 32, a determination module 34, and a detection module 36, which are described in detail below.
An acquisition module 32 for acquiring heart rate data and acceleration data; a determining module 34, connected to the acquiring module 32, for determining a heart rate characteristic according to the heart rate data and determining an acceleration characteristic according to the acceleration data; and a detecting module 36 connected to the determining module 34 for detecting a sleep state according to the heart rate characteristic and the acceleration characteristic, wherein the sleep state includes a sleep start, a sleep end, a deep sleep stage, and a light sleep stage.
By the device, the acquisition module 32 is adopted to acquire heart rate data and acceleration data; the determining module 34 determines heart rate characteristics from the heart rate data and acceleration characteristics from the acceleration data; the detection module 36 detects the sleep state according to the heart rate characteristics and the acceleration characteristics, wherein the sleep state comprises a sleep start mode, a sleep end mode, a deep sleep stage mode and a shallow sleep stage mode, the heart rate characteristics and the acceleration characteristics are determined according to the heart rate data and the acceleration data, the sleep state is detected according to the heart rate characteristics and the acceleration characteristics, the purpose of detecting the sleep state at any time is achieved, the technical effect of all-day sleep detection is achieved, and the technical problems that sleep monitoring in the related art has certain limitation and all-day sleep detection is not supported are solved.
According to another aspect of the embodiments of the present invention, there is also provided a storage medium, where the storage medium includes a stored program, and when the program runs, the apparatus where the storage medium is located is controlled to execute the sleep state detection method of any one of the above.
According to another aspect of the embodiments of the present invention, there is also provided a processor, configured to execute a program, where the program executes to perform the sleep state detection method of any one of the above.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
In the above embodiments of the present invention, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.
In the embodiments provided in the present application, it should be understood that the disclosed technology can be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units may be a logical division, and in actual implementation, there may be another division, for example, multiple units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, units or modules, and may be in an electrical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic or optical disk, and other various media capable of storing program codes.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.

Claims (15)

1. A sleep state detection method, comprising:
acquiring heart rate data and acceleration data;
determining heart rate characteristics according to the heart rate data, and determining acceleration characteristics according to the acceleration data;
detecting sleep states according to the heart rate characteristics and the acceleration characteristics, wherein the sleep states comprise a sleep start stage, a sleep end stage, a deep sleep stage and a shallow sleep stage.
2. The method of claim 1, wherein acquiring the heart rate data and the acceleration data comprises:
detecting whether the wearable equipment is in a wearing state;
and under the condition that the wearable equipment is in a wearing state, acquiring heart rate data and acceleration data recorded by the wearable equipment.
3. The method of claim 1, wherein determining the heart rate characteristic from the heart rate data and determining the acceleration characteristic from the acceleration data comprises:
sampling the heart rate data, determining first sampling data, and taking the first sampling data as the heart rate characteristic;
and/or the presence of a gas in the gas,
and sampling the acceleration data, determining second sampling data, and taking the second sampling data as the acceleration characteristic.
4. The method of claim 1, wherein determining the heart rate characteristic from the heart rate data and determining the acceleration characteristic from the acceleration data comprises:
performing data extraction on the heart rate data according to an extraction window with preset duration;
converting the heart rate data within the extraction window into a finite number of first discrete features, wherein the first discrete features include at least one of: heart rate change rate;
taking the first discrete feature as the heart rate feature;
and/or the presence of a gas in the gas,
performing data extraction on the acceleration data according to an extraction window with preset duration;
converting the acceleration data within the extraction window into a finite number of second discrete features, wherein the second discrete features include at least one of: activity intensity, activity duration, inactivity duration, activity and inactivity switching times;
and taking the second discrete characteristic as the acceleration characteristic.
5. The method of claim 1, wherein determining the heart rate characteristic from the heart rate data and determining the acceleration characteristic from the acceleration data comprises:
performing data extraction on the heart rate data according to an extraction window with preset duration;
converting the difference and variance of the heart rate data extracted from a plurality of continuous extraction windows into a first conversion characteristic, wherein the first conversion characteristic comprises heart rate change rate, and the extraction windows extract the heart rate data at a preset frequency;
taking the first conversion characteristic as the heart rate characteristic;
and/or the presence of a gas in the gas,
performing data extraction on the acceleration data according to an extraction window with preset duration;
converting the difference and variance of the acceleration data extracted from a plurality of continuous extraction windows into second conversion characteristics, wherein the second conversion characteristics comprise an active state, an active time length, a static state, a static time length and the number of active and static switching times, and the extraction windows extract the acceleration data at a preset frequency;
and taking the second conversion characteristic as the acceleration characteristic.
6. The method of any one of claims 3 to 5, wherein the sleep state further comprises: the sleep stages, secondary sleep, in which sleep is in, begin; detecting a sleep state according to the heart rate characteristic and the acceleration characteristic, comprising:
judging whether the current sleep state is in, carrying out staged detection on a plurality of sleep stages of the sleep state under the condition of the sleep state to determine the sleep stage, and then carrying out detection of sleep completion;
under the condition that the sleep state is not achieved, whether the time interval from the last sleep exceeds a preset interval threshold value or not is judged, and under the condition that the time interval exceeds the preset interval threshold value, the sleep start is detected;
and under the condition that the time interval does not exceed the preset interval threshold, detecting the start of the secondary sleep.
7. The method of claim 6, wherein detecting a sleep state based on the heart rate characteristic and the acceleration characteristic comprises: detecting whether the sleep state is sleep start or not according to the heart rate characteristics and the acceleration characteristics;
detecting sleep onset of the sleep state based on the heart rate characteristic and the acceleration characteristic comprises: inputting the heart rate characteristic or the acceleration characteristic into a first logic condition group, and outputting a detection result by the first logic condition group, wherein the first logic condition group comprises a plurality of first logic conditions, and wherein the first logic condition group comprises at least one of the following conditions: whether the static duration reaches a first preset static duration or not, whether the activity duration is lower than a first preset activity duration or not, whether the activity amplitude is smaller than a first preset amplitude or not and whether the activity intensity is lower than a first preset activity intensity or not;
alternatively, the first and second electrodes may be,
detecting whether the sleep state is sleep onset according to the heart rate characteristic and the acceleration characteristic comprises: inputting the heart rate characteristics or the acceleration characteristics into a first recognition model, and outputting a first detection result by the first recognition model, wherein the first recognition model is formed by training a plurality of groups of training data, and each group of training data comprises the heart rate characteristics or the acceleration characteristics and a corresponding first detection result.
8. The method of claim 6, wherein detecting a plurality of sleep stages in which a sleep state is located based on the heart rate characteristic and the acceleration characteristic further comprises: detecting a plurality of sleep stages in a sleep state in stages according to the heart rate characteristics and the acceleration characteristics, wherein the sleep stages include at least one of: a light sleep stage, a deep sleep stage;
staging detection of multiple sleep stages in a sleep state according to the heart rate signature and the acceleration signature comprises: and judging whether the heart rate characteristic and the acceleration characteristic are in a preset threshold range or not, and determining the sleep stage to be the sleep stage corresponding to the preset threshold range under the condition of being in the preset threshold range.
9. The method of claim 8, wherein prior to detecting the plurality of sleep stages in the sleep state periodically based on the heart rate characteristic and the acceleration characteristic, further comprising:
predicting the preset threshold value ranges corresponding to different sleep stages in the current sleep state according to the historical record of the last sleep state, wherein the historical record comprises heart rate characteristics and acceleration characteristics of the user in a plurality of different sleep stages;
alternatively, the first and second electrodes may be,
and under the condition of initialization, emptying a historical record, and predicting the preset threshold value ranges corresponding to different sleep stages in the current sleep state according to the prior value given by the initialization, wherein the historical record comprises heart rate characteristics and acceleration characteristics of the user in a plurality of different sleep stages.
10. The method of claim 6, wherein detecting a sleep state based on the heart rate characteristic and the acceleration characteristic further comprises: detecting the onset of a second sleep according to the heart rate characteristic and the acceleration characteristic;
detecting onset of a secondary sleep based on the heart rate characteristic and the acceleration characteristic comprises: inputting the heart rate characteristic or the acceleration characteristic into a second logic condition group, and outputting a detection result by the second logic condition group, wherein the second logic condition group comprises a plurality of second logic conditions; the second logic condition group comprises whether the static time length reaches a second preset static time length or not, whether the activity time length is lower than a second preset activity time length or not, whether the activity amplitude is smaller than a second preset amplitude or not and whether the activity intensity is lower than a second preset activity intensity or not;
alternatively, the first and second electrodes may be,
detecting onset of a secondary sleep based on the heart rate characteristic and the acceleration characteristic comprises: and inputting the heart rate characteristics or the acceleration characteristics into a second recognition model, and outputting a second detection result by the second recognition model, wherein the second recognition model is formed by training a plurality of groups of training data, and each group of training data comprises the heart rate characteristics or the acceleration characteristics of the second sleep and the corresponding second detection result.
11. The method of claim 10, wherein the heart rate signature and acceleration signature in the sleep state prior to the wake session corresponding to the secondary sleep are deleted prior to detecting the onset of secondary sleep based on the heart rate signature and the acceleration signature.
12. The method of claim 6, wherein detecting a sleep state based on the heart rate characteristic and the acceleration characteristic comprises: detecting whether the sleep state is sleep termination according to the heart rate characteristics and the acceleration characteristics;
detecting the end of sleep in the sleep state based on the heart rate characteristic and the acceleration characteristic comprises: inputting the heart rate characteristic or the acceleration characteristic into a third logic condition set, and outputting a detection result by the third logic condition set, wherein the third logic condition set comprises a plurality of third logic conditions, and wherein the third logic condition set comprises at least one of the following conditions: whether the activity duration reaches a third preset activity duration, the activity amplitude reaches a third preset amplitude, and whether the activity intensity reaches a third preset activity intensity;
alternatively, the first and second electrodes may be,
detecting whether the sleep state is an end of sleep according to the heart rate characteristic and the acceleration characteristic comprises: and inputting the heart rate characteristics or the acceleration characteristics into a third recognition model, and outputting a third detection result by the third recognition model, wherein the third recognition model is formed by training a plurality of groups of training data, and each group of training data comprises the heart rate characteristics or the acceleration characteristics and a corresponding third detection result.
13. A sleep state detection apparatus, comprising:
the acquisition module is used for acquiring heart rate data and acceleration data;
the determining module is used for determining heart rate characteristics according to the heart rate data and determining acceleration characteristics according to the acceleration data;
the detection module is used for detecting sleep states according to the heart rate characteristics and the acceleration characteristics, wherein the sleep states comprise a sleep start stage, a sleep end stage, a deep sleep stage and a shallow sleep stage.
14. A storage medium comprising a stored program, wherein the apparatus in which the storage medium is located is controlled to perform the sleep state detection method according to any one of claims 1 to 12 when the program is executed.
15. A processor configured to run a program, wherein the program is configured to execute the sleep state detection method according to any one of claims 1 to 12 when running.
CN201911089684.6A 2019-11-08 2019-11-08 Sleep state detection method and device Pending CN110710962A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911089684.6A CN110710962A (en) 2019-11-08 2019-11-08 Sleep state detection method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911089684.6A CN110710962A (en) 2019-11-08 2019-11-08 Sleep state detection method and device

Publications (1)

Publication Number Publication Date
CN110710962A true CN110710962A (en) 2020-01-21

Family

ID=69214945

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911089684.6A Pending CN110710962A (en) 2019-11-08 2019-11-08 Sleep state detection method and device

Country Status (1)

Country Link
CN (1) CN110710962A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112494192A (en) * 2020-12-24 2021-03-16 喜临门家具股份有限公司 Spine protection system
CN113229808A (en) * 2021-05-19 2021-08-10 珠海格力电器股份有限公司 Method and device for judging whether a person is in bed or not, storage medium and electronic equipment
CN114532992A (en) * 2022-03-23 2022-05-27 深圳市爱都科技有限公司 Method, device and system for detecting snooze state and computer readable storage medium

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105640508A (en) * 2016-03-30 2016-06-08 安徽华米信息科技有限公司 Realtime sleep monitoring method and device and intelligent wearable equipment
CN105902257A (en) * 2016-06-27 2016-08-31 安徽华米信息科技有限公司 Sleep state analysis method and device and intelligent wearable equipment
CN106108843A (en) * 2016-06-17 2016-11-16 美的集团股份有限公司 A kind of user of determination sleeps the method and system of duration
CN107872965A (en) * 2016-05-09 2018-04-03 倍灵科技有限公司 Wearable device and its method for health care
CN107890339A (en) * 2017-11-09 2018-04-10 常熟理工学院 A kind of sleep stage detection method and wearable sleep stage detection means
CN107890340A (en) * 2017-12-06 2018-04-10 上海斐讯数据通信技术有限公司 A kind of sleep monitor method and system
CN110151136A (en) * 2019-05-28 2019-08-23 深圳如一探索科技有限公司 It has ready conditions with reference to heart rate sleep state monitoring method, device, equipment and medium

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105640508A (en) * 2016-03-30 2016-06-08 安徽华米信息科技有限公司 Realtime sleep monitoring method and device and intelligent wearable equipment
CN107872965A (en) * 2016-05-09 2018-04-03 倍灵科技有限公司 Wearable device and its method for health care
CN106108843A (en) * 2016-06-17 2016-11-16 美的集团股份有限公司 A kind of user of determination sleeps the method and system of duration
CN105902257A (en) * 2016-06-27 2016-08-31 安徽华米信息科技有限公司 Sleep state analysis method and device and intelligent wearable equipment
CN107890339A (en) * 2017-11-09 2018-04-10 常熟理工学院 A kind of sleep stage detection method and wearable sleep stage detection means
CN107890340A (en) * 2017-12-06 2018-04-10 上海斐讯数据通信技术有限公司 A kind of sleep monitor method and system
CN110151136A (en) * 2019-05-28 2019-08-23 深圳如一探索科技有限公司 It has ready conditions with reference to heart rate sleep state monitoring method, device, equipment and medium

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112494192A (en) * 2020-12-24 2021-03-16 喜临门家具股份有限公司 Spine protection system
CN112494192B (en) * 2020-12-24 2023-11-07 喜临门家具股份有限公司 Ridge protection system
CN113229808A (en) * 2021-05-19 2021-08-10 珠海格力电器股份有限公司 Method and device for judging whether a person is in bed or not, storage medium and electronic equipment
CN114532992A (en) * 2022-03-23 2022-05-27 深圳市爱都科技有限公司 Method, device and system for detecting snooze state and computer readable storage medium
CN114532992B (en) * 2022-03-23 2024-04-02 深圳市爱都科技有限公司 Method, device and system for detecting nap state and computer readable storage medium

Similar Documents

Publication Publication Date Title
CN110710962A (en) Sleep state detection method and device
US10912495B2 (en) Activity recognition
CN104720746A (en) Sleeping stage determination method and system
CN104720748B (en) A kind of sleep stage determines method and system
CN110709940B (en) Method, system and medium for predicting sensor measurement quality
CN104200234A (en) Human body action modeling and recognizing method
CN112401838B (en) Method for detecting sleep state by wearable device and wearable device
CN108523899A (en) Monitoring method, storage medium in sleep procedure from bed state and device
CN111657855B (en) Sleep evaluation and sleep awakening method and device and electronic equipment
CN109620158B (en) Sleep monitoring method, intelligent terminal and storage device
CN106725325A (en) Analysis of sleeping quality method and analytical equipment
CN108095694A (en) State monitoring method, device and the wearable device of wearable device
CN109074057A (en) Operating personnel&#39;s managing device
CN115137308A (en) Method for improving accuracy of in-out sleep detection in sleep algorithm of intelligent wearable device
CN111473797A (en) Motion state detection method and device
CN106821336A (en) Sleep monitor method and system
CN108348157A (en) Utilize the heart rate detection of multipurpose capacitive touch sensors
CN106730232A (en) A kind of intelligent awakening method and system
CN110584675B (en) Information triggering method and device and wearable device
CN116803337B (en) Music sleep assisting method, device and equipment based on sleep monitoring
CN106333674B (en) Sleep cycle detection method and system in sleep state analysis
CN112274902B (en) Swimming data acquisition method and device and control method of swimming data acquisition equipment
CN106361327B (en) Waking state detection method and system in sleep state analysis
CN106175755B (en) Sleep state detector for sleep state analysis
CN113520339B (en) Sleep data validity analysis method and device and wearable device

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20200121

RJ01 Rejection of invention patent application after publication