WO2023284814A1 - 一种基于深度学习算法的电动床控制方法、系统和计算机程序 - Google Patents

一种基于深度学习算法的电动床控制方法、系统和计算机程序 Download PDF

Info

Publication number
WO2023284814A1
WO2023284814A1 PCT/CN2022/105657 CN2022105657W WO2023284814A1 WO 2023284814 A1 WO2023284814 A1 WO 2023284814A1 CN 2022105657 W CN2022105657 W CN 2022105657W WO 2023284814 A1 WO2023284814 A1 WO 2023284814A1
Authority
WO
WIPO (PCT)
Prior art keywords
intervention
snoring
electric bed
period
deep learning
Prior art date
Application number
PCT/CN2022/105657
Other languages
English (en)
French (fr)
Inventor
单华锋
张建炜
丁少康
李松
Original Assignee
麒盛科技股份有限公司
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 麒盛科技股份有限公司 filed Critical 麒盛科技股份有限公司
Publication of WO2023284814A1 publication Critical patent/WO2023284814A1/zh

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61F—FILTERS IMPLANTABLE INTO BLOOD VESSELS; PROSTHESES; DEVICES PROVIDING PATENCY TO, OR PREVENTING COLLAPSING OF, TUBULAR STRUCTURES OF THE BODY, e.g. STENTS; ORTHOPAEDIC, NURSING OR CONTRACEPTIVE DEVICES; FOMENTATION; TREATMENT OR PROTECTION OF EYES OR EARS; BANDAGES, DRESSINGS OR ABSORBENT PADS; FIRST-AID KITS
    • A61F5/00—Orthopaedic methods or devices for non-surgical treatment of bones or joints; Nursing devices; Anti-rape devices
    • A61F5/56—Devices for preventing snoring
    • A—HUMAN NECESSITIES
    • A47—FURNITURE; DOMESTIC ARTICLES OR APPLIANCES; COFFEE MILLS; SPICE MILLS; SUCTION CLEANERS IN GENERAL
    • A47C—CHAIRS; SOFAS; BEDS
    • A47C19/00—Bedsteads
    • A47C19/22—Combinations of bedsteads with other furniture or with accessories, e.g. with bedside cabinets
    • A—HUMAN NECESSITIES
    • A47—FURNITURE; DOMESTIC ARTICLES OR APPLIANCES; COFFEE MILLS; SPICE MILLS; SUCTION CLEANERS IN GENERAL
    • A47C—CHAIRS; SOFAS; BEDS
    • A47C20/00—Head -, foot -, or like rests for beds, sofas or the like
    • A47C20/04—Head -, foot -, or like rests for beds, sofas or the like with adjustable inclination
    • A47C20/041—Head -, foot -, or like rests for beds, sofas or the like with adjustable inclination by electric motors
    • A—HUMAN NECESSITIES
    • A47—FURNITURE; DOMESTIC ARTICLES OR APPLIANCES; COFFEE MILLS; SPICE MILLS; SUCTION CLEANERS IN GENERAL
    • A47C—CHAIRS; SOFAS; BEDS
    • A47C21/00—Attachments for beds, e.g. sheet holders, bed-cover holders; Ventilating, cooling or heating means in connection with bedsteads or mattresses
    • 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
    • A61B5/4803—Speech analysis specially adapted for diagnostic purposes
    • 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
    • A61B5/4809—Sleep detection, i.e. determining whether a subject is asleep or not
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
    • A61H23/00—Percussion or vibration massage, e.g. using supersonic vibration; Suction-vibration massage; Massage with moving diaphragms
    • A61H23/02—Percussion or vibration massage, e.g. using supersonic vibration; Suction-vibration massage; Massage with moving diaphragms with electric or magnetic drive
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
    • A61H2201/00—Characteristics of apparatus not provided for in the preceding codes
    • A61H2201/01—Constructive details
    • A61H2201/0119—Support for the device
    • A61H2201/0138—Support for the device incorporated in furniture
    • A61H2201/0142—Beds

Definitions

  • the present invention relates to a control method and system for an electric bed, in particular to a method and system for monitoring snoring based on a deep learning algorithm and controlling an electric bed.
  • No. CN103251388A Chinese patent application discloses a method for monitoring and preventing snoring.
  • the method classifies the collected audio signals into severe snoring signals and mild snoring signals and trains severe and mild GMM templates, which are collected by mobile phones
  • the snoring signal after extracting features, is brought into the GMM template for classification. If it is judged to be severe snoring, the snoring interval is judged, and an external stimulus is taken to intervene in snoring.
  • the existing technology cannot accurately know whether the snoring interferes with the user's normal sleep after the snoring intervention. In particular, sometimes the intervention may interfere with the user's deep sleep and reduce the user's sleep quality.
  • the present invention provides an electric bed control method based on deep learning and voice recognition technology to solve the technical problems identified by the applicant in the prior art, such as failure to obtain user feedback, possible interference with user sleep, and other technical problems.
  • some embodiments of the present application provide an electric bed control method based on a deep learning algorithm, which includes the steps of: receiving real-time monitored snoring data corresponding to the snoring sound from the smart device after the start of the intervention judgment period; Receive body movement frequency monitoring data;
  • One of the intervention level settings, one of the multiple intervention intensity settings, and the intervention mode setting drive at least one component of the electric bed to intervene in the generation of the snoring sound; wherein, the multiple intervention levels are respectively based on different The number of snoring data received in the intervention judgment period is determined.
  • the multiple intervention levels include a high intervention level determined according to the first number of snore data received within the intervention judgment period, and a high intervention level determined according to the first number of snore data received within the intervention judgment period.
  • the intervention level is adjusted to an intervention level with a higher number of snoring sounds.
  • said driving at least one component of the electric bed includes driving a head lift motor of the electric bed to raise the head deck by a first angle for a first period of time, and returning the head deck to the rest position after completion. flat position; wherein the lifting speed of the motor and the first time period are determined according to the intervention intensity setting.
  • the determination of the snoring event is resumed after the head deck is returned to the flat position.
  • the head board after the head board is restored to the flat position, it is judged whether the human body is out of the deep sleep state by monitoring the number of body movements and based on the number of body movements, and the judgment result is recorded.
  • the snoring intervention record is displayed on the interface of the smart terminal and/or the snoring audio is displayed, and a user feedback interface is provided to collect user feedback on the snoring intervention record.
  • the optimal control mode and the optimal intervention level are displayed on the interface before the start of the next sleep cycle.
  • the driving of at least one component of the electric bed includes driving the vibrating massage unit of the electric bed to vibrate for a second time period, wherein the vibration frequency of the vibration and the length of the second time period are determined according to the The intervention intensity setting is determined.
  • the driving of at least one component of the electric bed includes driving the vibrating massage unit of the electric bed to vibrate for a second time period; wherein the vibration frequency of the vibration and the length of the second time period are according to the The intervention intensity setting is determined; and, driving the head lifting motor of the electric bed to lift the head bed board by a first angle and continuing for a first period of time, wherein the speed of the motor lifting and the first period of time are set according to the intervention intensity Sure.
  • a deep learning algorithm-based electric bed control system which is characterized in that: it includes an embedded development board, and the development board is used as a computer program carrier to execute the computer program to achieve any of the above A new electric bed control method based on deep learning algorithm.
  • the method and system in this application provide a snoring intervention process based on the mobile phone APP, which can be set based on the threshold value of the mobile phone APP snoring intervention, and the snoring level can be determined and adjusted accordingly by adjusting the threshold within a monitoring period. In addition, it can complete the recommendation of the best intervention level and intervention method based on the snoring intervention of the mobile phone APP.
  • Fig. 1 is a flow schematic diagram of a snoring monitoring method and an intervention method based on a deep learning algorithm according to an embodiment of the present application;
  • Fig. 2 is a schematic flowchart of a snoring monitoring method and an intervention method based on a deep learning algorithm according to an embodiment of the present application;
  • Fig. 3 is a functional schematic diagram showing layer-by-layer convolution, pixel-by-pixel convolution and conventional convolution comparison
  • Fig. 4 is a schematic diagram of the sampling principle in some embodiments of the present application
  • Fig. 4 is a schematic flowchart of a snoring monitoring and intervention system according to some embodiments of the present application.
  • Fig. 5 is a schematic flowchart of a snoring monitoring and intervention system according to some embodiments of the present application.
  • the electric bed control method in this application can be based on such an application scenario, that is, an electric bed with a control unit, which can drive a specific position of the electric bed in response to a driving signal, such as at least the head board, waist board, foot board, etc. to change the user's sleeping position. posture, so as to intervene in the user's snoring sound.
  • a driving signal such as at least the head board, waist board, foot board, etc.
  • the drive signal may originate from a local control unit associated with the electric bed or from a remote control device coupled to the electric bed through the communication module.
  • the control unit or control device can receive monitoring signals and setting parameters from smart devices such as smart phones, and adjust the control signals accordingly.
  • the smart device can be used to receive a sleep cycle start signal from the control unit or the control device, and start monitoring the audio of the environment to judge whether there is snoring, and send the judgment result, such as snoring data, to the control unit or
  • the control device is used to determine whether the user is snoring.
  • the control unit or control device may receive the snoring sound data, and judge whether the user is snoring according to other data received from the electric bed, such as the number of body movements monitored by the body movement monitoring unit.
  • the smart device can be configured to be able to set the intervention parameters of the electric bed on the snoring behavior, such as the intervention level and intervention mode, and send the intervention parameters to the control unit or control device after establishing a connection with the control unit or control device It serves as the basis for the control unit or control device to provide the driving signal to the driving unit.
  • the above-mentioned functions of the smart device can be realized through the APP.
  • the APP can provide an intervention parameter setting interface to receive the user's setting and use the setting as the provided parameter of the control unit or control device.
  • the APP can monitor the audio signal of the environment and judge whether the audio line signal is a snoring sound through a known method or a specific method, and is configured to transmit the judgment result to the control unit or the control unit in real time if it is a snoring sound. equipment.
  • the control unit or control device may be any device with a processor and memory.
  • the controller can be directly or communicatively connected to electric beds and smart devices.
  • control method in the present application may be based on the monitoring results obtained by the snoring monitoring method of the deep learning algorithm.
  • the monitoring method can be the following specific monitoring method or any known monitoring method in the prior art.
  • Fig. 1 shows the flow chart of the embodiment of the snore monitoring method based on the exemplary deep learning algorithm according to the present invention, as shown in Fig. 1, the method comprises the following steps:
  • Step S101 collecting sound signals, wherein the sampling frequency can be set to 16KHz, and the sampling process will continue to run from the time the system is turned on, and will not stop until the system is actively turned off.
  • the sampling frequency can be set to 16KHz
  • Step S102 slice the collected audio signal to a specified length.
  • the length of 6 seconds is used, and the length of 3 seconds adopts the sliding window method, that is, whenever the continuously collected signal meets the new 6-second data When generated, audio slices are extracted from it.
  • Step S103 performing a mute algorithm analysis on the obtained fixed-length audio slice, and judging whether the audio segment contains sound;
  • Step S104 if the selected audio slice contains sound, perform acoustic feature extraction
  • Step S105 input the extracted acoustic features into the pre-trained deep neural network for deep feature extraction, and finally encode the audio segment into fixed-dimensional features, 128-dimensional features are used in this embodiment;
  • Step S106 input the obtained fixed-dimensional features into the fully connected layer network for final binary classification, and obtain whether the current audio slice is a snoring sound.
  • step S107 the above classification results are stored in a fixed-length queue, and then the total audio slice results in the queue are integrated and analyzed to determine whether there is a snoring event within the fixed time period.
  • the fixed-length queue may include the classification result and the audio slices that are determined not to contain sound in step S104.
  • a snoring event can be established according to the proportion of audio slices containing snoring in the above-mentioned fixed-length queue, and an intervention cycle can be started according to the number of occurrences of the snoring event to control the action or action sequence of the electric bed to intervene in the user's snoring.
  • audio can be continuously collected and the above-mentioned monitoring method can be carried out for monitoring. If there is no snoring event within a predetermined period of time, for example, within the intervention period, the intervention period will end, that is, the operation of the electric bed will be stopped. action.
  • a body movement monitoring unit may be provided for the electric bed, and the body movement monitoring unit may be, for example, a pressure sensor or a pressure sensor array to monitor the body movement data of the user on the electric bed, step S201.
  • the body motion monitoring unit can communicate with the controller of the electric bed to transmit the body motion data to the controller of the electric bed for judging whether there is body motion and recording the number of body motions. For example, whether there is body movement can be judged by the controller of the electric bed executing the judgment program, and the judgment result is sent to the electric bed control program executed for intervening snoring as the execution condition of whether to start the intervention.
  • the body motion data can be directly transmitted to the electric bed control program executed for snoring intervention, and a process in the program judges whether there is body motion and uses the judgment result as the execution condition of whether to start the intervention, step S202.
  • the number of body movements can be used as a condition of the electric bed control program of the present application.
  • a threshold can be set for the number of body movements, such as 20 body movements, and when it is judged that the number of body movements determined according to the body movement data within 30 minutes does not reach the threshold, the electric movement is allowed.
  • the start of the bed control program means that the user on the electric bed can be intervened according to the generation of the snoring event, step S203, step S204. If the set threshold is exceeded, the control of the electric bed is not activated even if there is a snoring event.
  • an intervention level may be set for the intervention, for example step S205 , the intervention level may include a low level S2051 , a medium level S2052 and a high level S2053 .
  • Low, medium, and high intervention levels can be determined according to the proportion of snoring events, as will be described in detail later.
  • an intervention mode can be set for the intervention, for example, step S206 in the figure.
  • the intervention can be, for example, starting the bedside motor of the electric bed as shown in FIG. 2 to raise the head of the bed at a certain angle, step S2061, or it can be, for example, starting the vibration massage unit of the electric bed to generate vibration, step S2062. Or a combination of lifting the head of the bed and shaking, step S2063.
  • Different intervention intensities can also be preset for each intervention mode, for example, as shown in detail in steps S2061, S2062, and S2063 and will be described in detail in subsequent descriptions.
  • the above queue can be emptied. Thereafter, the intervention state can be maintained for a period of time, and then it is monitored whether there is an audio slice of snoring in the monitoring period during the intervention period, and if there is an audio slice of snoring, the intervention is continued until there are no more audio slices.
  • the relevant algorithms of the present invention can be transplanted into embedded devices to run to obtain a snoring monitoring device, which can be combined with electric beds, pillows and other products to intervene in snoring behavior.
  • an ARM architecture embedded development board is used as the carrier of the algorithm and method of the present application, and a convolutional neural network is designed using a separate convolutional structure.
  • a convolutional neural network is designed using a separate convolutional structure.
  • the calculation amount of the convolution operation can be reduced to less than 1/9 of the original.
  • This structure can greatly reduce the amount of network parameters and calculations, and can be widely used in embedded smart devices such as mobile phones.
  • the control of the electric bed to realize the various parameters of the intervention of the user's snoring can be completed by the controller according to the automatic setting, or can be provided to the smart device, such as an APP deployed on the smart device, and receive the user's choice.
  • Setup is done by selecting an input, or by providing a setting suggestion to the smart device and confirming it with the user.
  • the intervention level can be set automatically, or the intervention level can be manually set through the APP, or the recommended intervention level can be provided on the APP for the user to confirm.
  • four intervention level options of "off", "high”, “medium” and “low” can be provided on the APP interface for selection.
  • the intervention level may correspond to the judgment of the snoring event.
  • the level "high” may correspond to a lower proportion of audio slices containing snoring in the judgment of snoring events.
  • a snoring event For example, if there are 5 audio slices containing snoring within the judgment period, a snoring event is generated;
  • the level "middle" corresponds to the medium proportion of audio slices containing snoring in the judgment of snoring events, for example, within the judgment period, 10 audio slices containing snoring are generated to generate snoring events;
  • a high proportion of audio slices containing snoring for example, 30 audio slices containing snoring within a determination period generates a snoring event.
  • the intervention method can be set automatically, or the intervention method can be manually set by using the APP, or the recommended intervention method can be provided on the APP for the user to confirm.
  • An APP interface can be provided for users to choose the control mode of the electric bed, that is, the intervention mode. For example, you can choose to lift the head of the electric bed for a period of time, or vibrate and massage for a period of time. Corresponding to these options, when the control conditions are met, a corresponding control instruction is sent to the controller of the electric bed.
  • the way of controlling the electric bed to achieve the intervention may include controlling the headboard motor of the electric bed to lift the headboard and keep it for 30 minutes, and the lifting angle may vary according to the bed type. For example, it can be set to lift 15 degrees, and then drive the headboard motor to restore the electric bed to a horizontal position.
  • the electric headboard Every time the electric headboard is lifted until it is flattened, it is set as an intervention cycle. After each intervention cycle, whether the human body is out of the deep sleep state is monitored and recorded.
  • the monitoring of the deep sleep state can be realized according to the above-mentioned body movement monitoring unit, or various methods and/or units that have been described in the prior art. If it is judged that the human body will leave the deep sleep state after one intervention period, the program will automatically downgrade the intervention level. ” to turn off the intervention.
  • the raising of the headboard may comprise different intervention intensities.
  • the intervention strength can be determined automatically, and can be provided on the APP interface for user selection, or automatically determined for user confirmation.
  • the intensity can include low intervention intensity, medium intervention intensity and high intervention intensity.
  • the electric bed can be set to low intervention intensity: the time to raise the head of the bed is 35 minutes, and the lifting speed is slow; in the middle intervention intensity: the time to raise the head of the bed is 25 minutes, The lifting speed is medium; at high intervention intensity: the time to lift the head of the bed is 15 minutes, and the lifting speed is fast.
  • the head lifting speed of different bed types is inconsistent, but there will be a fixed lifting speed as a benchmark.
  • the above-mentioned slow, medium and fast standards can be determined according to the fixed lifting speed of each bed type. For example, slow: about 1/5 of the fixed speed, medium: about 1/4, high: about 1/3.
  • slow about 1/5 of the fixed speed
  • medium about 1/4
  • high about 1/3.
  • speed of the motor is 7mm/s, by multiplying the above ratio, the specific rising speed value can be obtained.
  • the electric bed can be provided with the electric bed controller to provide units for intervention measures such as sound, light, smell, vibration, etc., and the electric bed controller can be configured to receive information from the mobile phone APP after the festival.
  • the electric bed controller can be configured to receive information from the mobile phone APP after the festival.
  • the vibration intervention can be realized by a vibration massage unit, which can be operatively coupled to the controller of the electric bed for controlled actions, for example, the vibration massage unit can be activated according to the setting of the intervention level and the snoring event to achieve the vibration intervention.
  • vibration intervention can include low intervention intensity, medium intervention intensity and high intervention intensity, for example, at low intervention intensity: vibration time 3 s, vibration frequency low; at medium intervention intensity: vibration time 6 seconds, vibration frequency medium; at high Intervention intensity: The vibration time is 9 seconds, and the vibration frequency is high.
  • a combined mode of head-raising intervention and vibration intervention may also be provided as an intervention manner.
  • the combination mode can be set by providing different intervention intensities, for example, in low intervention intensity: lifting time 35 minutes, lifting speed is slow, vibration time 3 seconds, vibration frequency is low; in medium intervention intensity: lifting time 25 minutes, the lifting speed is medium, the vibration time is 6 seconds, and the vibration frequency is medium; at high intervention intensity: the lifting time is 15 minutes, the lifting speed is fast, the vibration time is 9 seconds, and the vibration frequency is high.
  • the records of the intervention level, intervention mode, intervention intensity and intervention effect in multiple intervention cycles can be used to determine the intervention level and intervention mode of the next intervention cycle. For example, if each intervention period results in a decrease in the intervention level, ie, if a snoring event can still be detected in the next period after two "weak" level interventions, then the intervention level is increased by one level. For another example, if the snoring event can still be detected in the next cycle after adjusting the angle of the head board to 15 degrees twice in a row, then increase the elevation angle of the head board in the next intervention cycle.
  • the number of body movements detected by the body movement monitoring unit of the electric bed can be used to judge the user's sleep status into the following three categories:
  • the program automatically adjusts the snoring intervention level, intensity and mode currently set by the user.
  • the adjustment instructions are as follows.
  • the system When the user's sleep is disturbed after the intervention, the system will automatically reduce the intervention level and intervention intensity set by the current user, or if the highest value of the intervention intensity of a certain intervention method is still insufficient or the lowest value is still too high, the system will automatically reduce the user's sleep without disturbing the user's sleep. Adjust to other intervention methods; for example, when the maximum intensity of vibration intervention is still insufficient, start the intervention method of raising the head board, or when the intervention method of raising the head board at the maximum intensity is still insufficient, start the combination of raising the head board and vibration intervention blend mode.
  • the system When the user's snoring is not relieved after the intervention, the system will automatically increase the intervention intensity currently set by the user, or automatically adjust to other intervention methods without disturbing the user's sleep. During the intervention process, the system will record the relevant data change information and save it to the cloud database, and start the snoring intervention again according to the modified snoring intervention level and method.
  • the user When it is judged that the user is awake and out of bed, the user is guided to score and evaluate the snoring intervention. .
  • the scoring and evaluation can be displaying the intervention record on the APP at a fixed frequency or a preset frequency or according to the user's call, and displaying audio slices containing snoring for the user to view.
  • a user feedback interface can also be provided on the interface at the same time, for example, a selection object can be provided on the interface to allow the customer to select whether a snoring record is a false alarm. Relying on this method can further improve the accuracy of the snore recognition algorithm.
  • the mattress can be partially inflated to change the position of the head or body so as to achieve the anti-snoring effect by changing the user's sleeping position.
  • the mattress can be provided with a partial inflation function, and a partial inflation device can be provided for the mattress, that is, a partial location of the mattress can be inflated to change its surface shape, such as making the head of the mattress higher than other locations.
  • the partial inflatable device is connected with the controller of the electric bed, and when the electric bed controller receives a signal from the mobile phone APP to start snoring intervention, the inflatable cushion is partially inflated

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Veterinary Medicine (AREA)
  • Public Health (AREA)
  • Animal Behavior & Ethology (AREA)
  • Biomedical Technology (AREA)
  • Engineering & Computer Science (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Pathology (AREA)
  • Biophysics (AREA)
  • Physics & Mathematics (AREA)
  • Nursing (AREA)
  • Pulmonology (AREA)
  • Otolaryngology (AREA)
  • Anesthesiology (AREA)
  • Orthopedic Medicine & Surgery (AREA)
  • Vascular Medicine (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Epidemiology (AREA)
  • Pain & Pain Management (AREA)
  • Physical Education & Sports Medicine (AREA)
  • Rehabilitation Therapy (AREA)
  • Orthopedics, Nursing, And Contraception (AREA)
  • Invalid Beds And Related Equipment (AREA)

Abstract

本发明公开了基于深度学习算法的电动床控制方法，其包括以下步骤：在干预判断周期开始后自智能设备接收实时监测到的对应鼾声的鼾声数据；自电动床控制器接收体动次数监测数据；判断所述体动次数监测数据是否大于预设阈值，如大于所述预设阈值则在干预周期内根据在所述干预判断周期内接收到的鼾声数据的次数基于自所述智能设备接收的多个干预等级中的一个以及干预方式设置参数驱动所述电动床的至少一个部件动作以干预所述鼾声的产生。本申请还公开了对应的控制系统和计算机程序。本申请中的方法和系统提供了基于手机APP的打鼾干预过程，可基于手机APP打鼾干预的阈值设置，在一个监测周期内通过调整阈值来实现打鼾等级判定并进行相应调整。

Description

一种基于深度学习算法的电动床控制方法、系统和计算机程序 技术领域
本发明涉及电动床的控制方法和系统，尤其涉及一种基于深度学习算法的鼾声监测和电动床的控制方法和系统。
背景技术
据不完全统计，我国打鼾人群占比近两成，打鼾严重者甚至引发阻塞性呼吸暂停，严重影响身体健康。
第CN103251388A号中国专利申请公开了一种对鼾声进行监测并进行防治的方法，该方法对采集的音频信号分类为重度鼾声信号和轻度鼾声信号并训练出重度和轻度GMM模板，通过手机采集鼾声信号，提取特征后，带入GMM模板进行分类，如果判定为重度鼾声，则判断鼾声间隔，并采取外部刺激干预打鼾。
第CN105193566号中国专利申请公开了一种抑制打鼾的方法和智能床，其中公开了获取房内声音，判断房内声音是否大于预设的分贝值从而判断是否为鼾声并在判断出是鼾声后对应的抬升床板角度，以此抑制打鼾。
申请人发现，现有技术中的针对用户打鼾时的电动床的控制方法均无法得出适合特定用户的干预方式，因其并不获得来自用户的对干预操作的反馈。
同时现有技术在鼾声干预后，无法准确得知是否干扰用户的正常睡眠，尤其是，有的时候干预可能会干扰用户的深度睡眠，降低用户的睡眠质量。
发明内容
本发明提供一种基于深度学习和语音识别技术的电动床控制方法，以解决现有技术中存在的申请人已经识别的诸如无法获得用户反馈，可能干扰用户睡眠等技术问题。
为此，本申请的一些实施例提供了基于深度学习算法的电动床控制方法，其包括步骤：在干预判断周期开始后自智能设备接收实时监测到的对 应鼾声的鼾声数据；自电动床控制器接收体动次数监测数据；
判断所述体动次数监测数据是否大于预设阈值，如大于所述预设阈值则在干预周期内根据在所述干预判断周期内接收到的鼾声数据的次数基于自所述智能设备接收的多个干预等级设置中的一个、多个干预强度设置中的一个以及干预方式设置驱动所述电动床的至少一个部件动作以干预所述鼾声的产生；其中，所述多个干预等级分别根据不同的所述干预判断周期内接收到的鼾声数据的次数确定。
在一些实施例中，所述多个干预等级包括根据所述干预判断周期内接收到的鼾声数据的第一次数确定的高干预等级、根据所述干预判断周期内接收到的鼾声数据的第二次数确定的中干预等级，以及根据所述干预判断周期内接收到的鼾声数据的第三次数确定的低干预等级；其中，所述第一次数小于所述第二次数小于所述第三次数。
在一些实施例中，在所述驱动所述电动床的至少一个部件动作开始后，监测并记录在所述干预周期内的人体是否脱离深度睡眠状态。
在一些实施例中，当判断人体在所述干预周期内脱离深度睡眠状态，则将干预等级调整为具有较高鼾声次数的干预等级。
在一些实施例中，所述驱动电动床的至少一个部件包括驱动电动床的头部抬升电机将头部床板抬升第一角度并持续第一时间段，并在完成后将头部床板恢复到放平位置；其中所述电机抬升的速度以及所述第一时间段根据所述干预强度设置确定。
在一些实施例中，在将所述头部床板恢复到放平位置后重新开始所述鼾声事件的判断。
在一些实施例中，在将所述头部床板恢复到放平位置后，通过监测所述体动次数并基于体动次数判断人体是否脱离深度睡眠状态，并且记录判断结果。
在一些实施例中，在多个干预周期后，在所述智能终端的界面上显示鼾声干预记录并且/或者展示鼾声音频，同时提供用户反馈界面以采集用户对鼾声干预记录的反馈。
在一些实施例中，基于所述鼾声干预记录、所述用户对鼾声干预记录 的反馈中的至少一项在下一个睡眠周期开始前向所述界面显示最佳控制方式和最佳干预等级。
在一些实施例中，所述驱动电动床的至少一个部件包括驱动电动床的震动按摩单元震动并持续第二时间段，其中所述震动的震动频率以及所述第二时间段的长短根据所述干预强度设置确定。
在一些实施例中，所述驱动电动床的至少一个部件包括驱动电动床的震动按摩单元震动并持续第二时间段；其中所述震动的震动频率以及所述第二时间段的长短根据所述干预强度设置确定；以及，驱动电动床的头部抬升电机将头部床板抬升第一角度并持续第一时间段，其中所述电机抬升的速度以及所述第一时间段根据所述干预强度设置确定。
本申请的另一些实施例提供了一种基于深度学习算法的电动床控制系统，其特征在于：其包括嵌入式开发板，所述开发板作为计算机程序载体，执行所述计算机程序以实现上述任意一项的基于深度学习算法的电动床控制方法。
本申请的另一些实施例提供了计算机程序，包括用于当在计算机系统上执行所述计算机程序时执行根据以上任一方法的所有步骤的指令。
本申请中的方法和系统提供了基于手机APP的打鼾干预过程，可基于手机APP打鼾干预的阈值设置，在一个监测周期内通过调整阈值来实现打鼾等级判定并进行相应调整。另可基于手机APP打鼾干预完成最佳干预等级和干预方式的推荐。
附图说明
图1是根据本申请的实施例的基于深度学习算法的鼾声监测方法与干预方法的流程原理图；
图2是根据本申请的实施例的基于深度学习算法的鼾声监测方法与干预方法的流程示意图；
图3是显示逐层卷积、逐像素卷积和常规卷积比对的功能原理图；图4是本申请的一些实施例中的采样原理的示意图；
图4是本申请的一些实施例的鼾声监测和干预系统的流程示意图；
图5是本申请的一些实施例的鼾声监测和干预系统的流程示意图。
具体实施方式
下面结合附图对本申请的具体实施方式进行详细说明。
本申请中的电动床控制方法可以基于这样一个应用场景，即具有控制单元的电动床，其能够响应驱动信号驱动电动床的特定位置，例如至少头板、腰板、脚板等以改变使用者的睡姿，从而对使用者的鼾声进行干预。
所述驱动信号可以来源于一个与电动床相关联的本地的控制单元或者来源于通过通信模块耦合于电动床的远程的控制设备。所述控制单元或控制设备可以接收来自例如智能手机等智能设备的监测信号以及设置参数，并对应地调整所述控制信号。例如，所述智能设备可以用于从控制单元或控制设备接收睡眠周期开始的信号，并开始监测环境的音频以判断是否有鼾声，并将判断结果，例如鼾声数据，发送给所述控制单元或控制设备用于判断使用者是否在打鼾。所述控制单元或控制设备可以接收上述鼾声数据，并根据从电动床接收的其他数据，例如通过体动监测单元监测到的体动次数来判断使用者是否在打鼾。可以将智能设备配置为能够设置电动床对打鼾行为的干预参数，例如干预等级和干预方式，并在与所述控制单元或控制设备建立连接后将该干预参数发送至所述控制单元或控制设备以作为控制单元或控制设备向驱动单元提供驱动信号的依据。
在这里智能设备的上述功能可以通过APP实现。例如，该APP在运行后可以提供干预参数的设置界面以接收使用者的设置并将该设置作为提供的所述控制单元或控制设备的参数。该APP可以在运行后监测环境的音频信号并通过已知的方法或者特定的方法判断该音频线信号是否为鼾声，并被配置为如果为鼾声则将该判断结果实时地传送给控制单元或控制设备。
所述控制单元或控制设备可以是任何具有处理器以及存储器的设备。器可以直接或通信地连接于电动床以及智能设备。
在一些实施例中，本申请中的控制方法可以基于深度学习算法的鼾声监测方法得到的监测结果。监测方法可以是以下特定的监测方法也可以是 任何现有技术中已知的监测方法。
图1示出了根据本发明的示例性的基于深度学习算法的鼾声监测方法的实施例的流程图，如图1所示，该方法包括以下步骤：
步骤S101，采集声音信号，其中可以将采样频率设置为16KHz，采样过程从系统开启时便持续运行，直到主动关闭时才会停止采集。
步骤S102，对采集到的音频信号进行切片到指定长度，本实施例采用6秒长度，其中3秒的长度采用滑窗的方式，即每当持续不断地采集的信号中满足新的6秒数据产生时，就从中提取音频切片。
步骤S103，对得到的定长音频切片进行静音算法分析，判断音频片段中是否包含声音；
步骤S104，若所取音频切片中包含声音，则进行声学特征提取；
步骤S105，将提取好的声学特征输入到预先训练好的深度神经网络中进行深度特征提取，最终将音频段编码成固定维度的特征，本实施例中采用128维度特征；
步骤S106，将得到的固定维度的特征输入到全连接层网络中进行最终的二分类，得到当前音频切片是否为鼾声。
步骤S107，将上面分类的结果存入固定长度队列中，然后对队列中总的音频切片结果进行整合分析，判断在这个固定的时间段内是否存在鼾声事件。
固定长度队列中可以仅包括所述分类的结果。
固定长度队列中可以包括所述分类的结果以及步骤S104中判断为不包含声音的音频切片。
上述是否有鼾声的判断结果可以直接作为后续对电动床动作的控制依据。例如，可以根据上述固定长度队列中的包含鼾声的音频切片的比例来建立鼾声事件，并根据该鼾声事件发生的次数来启动干预周期以控制电动床的动作或动作序列以干预使用者的打鼾。
在电动床动作的过程中，可以继续采集音频并执行上述监测方法进行监测，如果在预定的时间内，例如在干预周期内时间段没有发生鼾声事件，则结束干预周期，即则停止电动床的动作。
为了更加准确地对是否处于打鼾状态进行判断，可以为干预的判断附加其他条件。如图2所示。
例如可以为电动床设置体动监测单元，体动监测单元例如可以为压力传感器或压力传感器阵列以监测电动床上的用户的体动数据，步骤S201。体动监测单元可以与电动床的控制器通信连接以将所述体动数据传输至电动床的控制器用于判断是否存在体动并记录体动次数。例如是否有体动可以有电动床的控制器执行判断程序进行判断，并将判断结果传送给为干预打鼾而执行的电动床控制程序作为是否启动干预的执行条件。或者，可以将体动数据直接传输给为干预打鼾而执行的电动床控制程序，由该程序中的一个进程判断是否存在体动并将判断结果作为是否启动干预的执行条件，步骤S202。
例如，体动的次数可以作为本申请的电动床控制程序的条件。例如，可以为体动次数设置阈值，例如20次体动，且当判断单位时间内，例如30分钟内根据所述体动数据确定的体动次数未达到所述阈值时，则允许所述电动床控制程序的启动，即可以根据鼾声事件的产生而对电动床上的用户进行干预，步骤S203、步骤S204。如果超过设置的阈值，则不启动对电动床的控制即便存在鼾声事件。
如图5所示，可以为干预设置干预等级，例如步骤S205，干预等级可以包括低等级S2051、中等级S2052和高等级S2053。低、中、高干预等级可以根据鼾声事件占比确定，如后续详述。
又如图5所示，在确定干预等级后，或者同时，可以为干预设置干预方式，例如图步骤S206。所述干预可以例如图2所示的启动电动床的床头电机抬起床头一定角度，步骤S2061，也可以是的例如启动电动床具有的震动按摩单元产生震动，步骤S2062。或者是抬起床头和震动的组合，步骤S2063。
还可以分别为每个干预方式预设不同的干预强度，例如步骤S2061、S2062、S2063中所详细示出的并将在后续描述中详细介绍。
执行上述干预步骤后，可以清空上述队列。此后可以保持干预状态一段时间，然后监测干预的时间段中的监测周期内是否有鼾声的音频切片， 如果有鼾声音频切片则继续保持干预，直至不再有音频切片。
进一步地，本发明相关算法均可移植到嵌入式设备中运行，来得到一款鼾声监测设备，并结合电动床、枕头等产品类进行打鼾行为干预。
本实施例采用ARM架构嵌入式开发板来作为本申请的算法、方法的载体，采用分离式卷积结构来设计卷积神经网络，分离式卷积神经网络和传统卷积结构的对比如图3所示，据图可以得出如下结论：
原始的卷积结构计算量为
C ori＝K*K*N*M*F out*F out        (2)
分离式卷积结构计算量为
C new＝N*M*F out*F out+K*K*M*F out*F out       (3)
与原来的卷积相比减少的计算量为
Figure PCTCN2022105657-appb-000001
也就是说如果原来卷积核大小为3*3，那么卷积操作的计算量能降到原来的1/9以下。这种结构可以大幅较少网络参数量和计算量，可广泛应用于手机等嵌入式智能设备中。
根据上述判断启动的对电动床的控制以实现对用户打鼾的干预的各种参数可以由控制器根据自动设置完成、也可以通过向智能设备，例如向智能设备部署的APP提供选择并接收用户的选择输入来设置完成，或者通过向智能设备提供设置建议并经用户确认而设置完成。
例如可以自动设置干预等级，也可以利用APP手动设置干预等级，或者在APP上提供推荐的干预等级后供用户确认。例如，可以在APP界面提供“关闭”、“高”、“中”、“低”四个干预等级选项供选择。干预等级可以对应鼾声事件的判断，例如可以将等级“高”对应鼾声事件判断中的较低的包含鼾声的音频切片比例，例如在判断周期内有5个包含鼾声的音频切片则生成鼾声事件；将等级“中”对应鼾声事件判断中的中等的包含鼾声的音频切片比例，例如在判断周期内有10个包含鼾声的音频切片则生成鼾声事件；将等级“低”对应鼾声事件判断中的较高的包含鼾声的音频切片比例，例如在判断周期内有30个包含鼾声的音频切片则生成鼾声事件。还可以选择“关闭”关闭干预功能，则系统仅记录包含鼾声的音频切 片及/或仅判断鼾声事件而不启动干预。
还例如可以自动设置干预方式，也可以利用APP手动设置干预方式，或者在APP上提供推荐的干预方式后供用户确认。可以提供APP界面供用户选择电动床的控制方式，即干预方式。例如可以选择上述的抬起电动床的床头一段时间，或者震动按摩一段时间。对应这些选项，当满足控制条件后，则向电动床的控制器发送对应的控制指令。
控制电动床以实现干预的方式可以包括控制电动床的头板电机抬升头板并保持30分钟，抬升角度可以根据床型的不同而不同。例如，可以设置为抬升15度，到时后驱动头板电机将电动床恢复到水平位置。
每次抬起直至放平电动床头板被设定为一个干预周期，每个干预周期结束后，监测人体是否脱离深度睡眠状态，并作记录。深度睡眠状态的监测可以根据上述体动监测单元实现，或者现有技术中已经描述的各种方法和/或单元实现。若判断人体一个干预周期结束后脱离深度睡眠状态，则程序自动对干预等级进行降级设置，例如用户当前设置的干预等级为“中”，则自动将干预等级降为“弱”，如果为“弱”则关闭干预。
所述抬起床头头板可以包括不同的干预强度。类似的，干预强度可以自动确定，可以提供在APP界面上供用户选择，或者自动确定后供用户确认。强度可以包括低干预强度、中干预强度和高干预强度，电动床可以被设置为在低干预强度：抬起床头时间35分钟，抬起速度缓慢；在中干预强度：抬起床头时间25分钟，抬起速度中等；在高干预强度：抬起床头时间15分钟，抬起速度快速。不同床型的头部抬起速度是不一致的，但都会有作为基准的固定抬升速度。因此，可以根据每一个床型的固定抬升速度确定上述缓慢、中等、快速的标准。例如缓慢：固定速度的1/5左右，中等：1/4左右，高：1/3左右。例如电机的速度为7mm/s，通过乘以上述比例，可以得到具体上升速度值。
除此之外可以为电动床提供与电动床控制器相连的例如声、光、气味、震动等干预措施的提供单元，并使得电动床控制器被配置为当节后到来自手机APP的因存在鼾声事件而发出的控制信号时，启动上述声、光、气味、震动等提供单元中的一个或多个，以通过改变灯光、声音、气味、震动等 环境因素，来达到止鼾效果。
以震动干预为例。震动干预可以通过震动按摩单元实现，震动按摩单元可以可操作地耦合于电动床的控制器以受控动作，例如可以根据干预等级的设置以及鼾声事件而启动震动按摩单元实现震动干预。类似的，震动干预可以包括低干预强度、中干预强度和高干预强度，例如，在低干预强度：震动时间3s，震动频率低；在中干预强度：震动时间6秒，震动频率中；在高干预强度：震动时间9秒，震动频率高。
作为替换或补充，在一些实施例中，还可以提供抬起头板干预和震动干预的组合模式作为干预方式。类似的，该组合模式可以通过提供不同的干预强度设置，例如在低干预强度：抬起时间35分钟，抬起速度缓慢，震动时间3秒，震动频率低；在中干预强度：抬起时间25分钟，抬起速度中等，震动时间6秒，震动频率中；在高干预强度：抬起时间15分钟，抬起速度快速，震动时间9秒，震动频率高。
可以将多个干预周期中的干预等级、干预方式、干预强度以及干预效果的记录用于确定下一干预周期的干预等级和干预方式。例如，如果每个干预周期都会造成干预等级降低，即，如果经过两次“弱”等级的干预仍能够在下一周期监测到鼾声事件，则将干预等级提升一个等级。又例如，如果连续两次通过调整头板的角度至15度角后，仍能够在下一周期监测到鼾声事件，则在下一个干预周期内增加头板的抬升角度。
一次干预结束后，则通过电动床设置体动监测单元检测到的体动次数可以判断用户睡眠状态主要为以下三类：
a)打鼾缓解，用户清醒，干扰用户睡眠；
b)打鼾缓解，用户未醒，不干扰用户睡眠；
c)打鼾没有缓解，不干扰用户睡眠。
根据出现的上述三类情况，记录用户当前打鼾干预方式、干预等级、干预强度、干预后用户状态。通过程序自动调整当前用户设置的打鼾干预等级、强度和方式，调整说明如下。
干预后打扰用户睡眠时，系统自动降低当前用户设置的干预等级以及干预强度，或在某一干预方式的干预强度的最高值仍不足或者最低值仍过 高后，不打扰用户睡眠的情况下自动调整为其他干预方式；例如在震动干预的最大强度仍不足时，则启动抬升头头板的干预方式，或者在最大强度的抬升头板的干预方式仍不足时，则启动抬升头板和震动干预的混合模式。
干预后用户打鼾未缓解时，系统自动提高当前用户设置的干预强度，或在不打扰用户睡眠的情况下自动调整为其他干预方式。干预的过程中，系统会记录相关数据变更信息保存至云端数据库，根据修改的打鼾干预等级和方式，再次启动打鼾干预。
当检测到干预后用户打鼾缓解且不打扰用户睡眠时，表明当前干预方式与等级为最佳的用户鼾声干预方式，记录相关数据到云端数据库，用户打鼾缓解后，再次进入用户睡眠检测，等待下一轮打鼾。
当判断用户清醒离床后，在引导用户对打鼾干预进行打分和评价。。
所述打分和评价可以为以固定频率或预设的频率或者根据用户的调取在APP上显示干预记录，并展示包含鼾声的音频切片，供用户查看。还可以同时在界面上提供用户的反馈接口，例如，可以在界面上提供选择对象，以允许客户选择一个鼾声的记录是否为误报。依靠这种方式可以进一步提高鼾声识别算法的准确性。
监测到鼾声事件后，除了上述可以通过抬升床板以进行干预外，可以通过使得床垫局部充气来改变头部或身体的位置从而通过改变用户睡姿达到止鼾效果。为此，可以为床垫提供局部充气功能，为床垫提供局部充气装置，即床垫的局部位置可以被充气以改变其表面形状，例如使得床垫的头部位置高于其他位置。该局部充气装置与电动床的控制器相连接，当电动床控制器收到来自手机APP的启动鼾声干预的信号后，则使得充气垫局部充气

Claims (13)

  1. 一种基于深度学习算法的电动床控制方法，其特征在于：包括步骤
    在干预判断周期开始后自智能设备接收实时监测到的对应鼾声的鼾声数据；自电动床控制器接收体动次数监测数据；
    判断所述体动次数监测数据是否大于预设阈值，如大于所述预设阈值则在干预周期内根据在所述干预判断周期内接收到的鼾声数据的次数基于自所述智能设备接收的多个干预等级设置中的一个、多个干预强度设置中的一个以及干预方式设置驱动所述电动床的至少一个部件动作以干预所述鼾声的产生；其中，所述多个干预等级分别根据不同的所述干预判断周期内接收到的鼾声数据的次数确定。
  2. 根据权利要求1的基于深度学习算法的电动床控制方法，其特征在于：所述多个干预等级包括根据所述干预判断周期内接收到的鼾声数据的第一次数确定的高干预等级、根据所述干预判断周期内接收到的鼾声数据的第二次数确定的中干预等级，以及根据所述干预判断周期内接收到的鼾声数据的第三次数确定的低干预等级；其中，所述第一次数小于所述第二次数小于所述第三次数。
  3. 根据权利要求2的基于深度学习算法的电动床控制方法，其特征在于：在所述驱动所述电动床的至少一个部件动作开始后，监测并记录在所述干预周期内的人体是否脱离深度睡眠状态。
  4. 根据权利要求3的基于深度学习算法的电动床控制方法，其特征在于：当判断人体在所述干预周期内脱离深度睡眠状态，则将干预等级调整为具有较高鼾声次数的干预等级。
  5. 根据权利要求1的基于深度学习算法的电动床控制方法，其特征在于：所述驱动电动床的至少一个部件包括驱动电动床的头部抬升电机将头部床板抬升第一角度并持续第一时间段，并在完成后将头部床板恢复到放平位置；其中所述电机抬升的速度以及所述第一时间段根据所述干预强度设置确定。
  6. 根据权利要求5的基于深度学习算法的电动床控制方法，其特征在于：在将所述头部床板恢复到放平位置后重新开始所述鼾声事件的判断。
  7. 根据权利要求5的基于深度学习算法的电动床控制方法，其特征在于：在将所述头部床板恢复到放平位置后，通过监测所述体动次数并基于体动次数判断人体是否脱离深度睡眠状态，并且记录判断结果。
  8. 根据权利要求7的基于深度学习算法的电动床控制方法，其特征在于：在多个干预周期后，在所述智能终端的界面上显示鼾声干预记录并且/或者展示鼾声音频，同时提供用户反馈界面以采集用户对鼾声干预记录的反馈。
  9. 根据权利要求9的基于深度学习算法的电动床控制方法，其特征在于：基于所述鼾声干预记录、所述用户对鼾声干预记录的反馈中的至少一项在下一个睡眠周期开始前向所述界面显示最佳控制方式和最佳干预等级。
  10. 根据权利要求1的基于深度学习算法的电动床控制方法，其特征在于：所述驱动电动床的至少一个部件包括驱动电动床的震动按摩单元震动并持续第二时间段，其中所述震动的震动频率以及所述第二时间段的长短根据所述干预强度设置确定。
  11. 根据权利要求1的基于深度学习算法的电动床控制方法，其特征在于：所述驱动电动床的至少一个部件包括驱动电动床的震动按摩单元震动并持续第二时间段；其中所述震动的震动频率以及所述第二时间段的长短根据所述干预强度设置确定；以及，驱动电动床的头部抬升电机将头部床板抬升第一角度并持续第一时间段，其中所述电机抬升的速度以及所述第一时间段根据所述干预强度设置确定。
  12. 一种基于深度学习算法的电动床控制系统，其特征在于：其包括嵌入式开发板，所述开发板作为计算机程序载体，执行所述计算机程序以实现上述权利要求1至11中任意一项的方法。
  13. 一种计算机程序，包括用于当在计算机系统上执行所述计算机程序时执行根据以上任一方法权利要求所述的方法的所有步骤的指令。
PCT/CN2022/105657 2021-07-15 2022-07-14 一种基于深度学习算法的电动床控制方法、系统和计算机程序 WO2023284814A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110799620.6 2021-07-15
CN202110799620.6A CN113599051A (zh) 2021-07-15 2021-07-15 一种基于深度学习算法的电动床控制方法、系统和计算机程序

Publications (1)

Publication Number Publication Date
WO2023284814A1 true WO2023284814A1 (zh) 2023-01-19

Family

ID=78337612

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/105657 WO2023284814A1 (zh) 2021-07-15 2022-07-14 一种基于深度学习算法的电动床控制方法、系统和计算机程序

Country Status (2)

Country Link
CN (1) CN113599051A (zh)
WO (1) WO2023284814A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113599051A (zh) * 2021-07-15 2021-11-05 麒盛科技股份有限公司 一种基于深度学习算法的电动床控制方法、系统和计算机程序
CN113599052A (zh) * 2021-07-15 2021-11-05 麒盛科技股份有限公司 一种基于深度学习算法的鼾声监测方法与系统以及相应的电动床控制方法和系统
CN114468655A (zh) * 2022-03-16 2022-05-13 浙江康贝德智能家居有限公司 一种自动调节床角度的电动床控制方法和系统

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2005152537A (ja) * 2003-11-28 2005-06-16 Yokogawa Electric Corp 制御システム
CN104188638A (zh) * 2014-09-09 2014-12-10 喜临门家具股份有限公司 一种睡眠保障系统及其控制方法
CN105517464A (zh) * 2013-03-14 2016-04-20 择舒公司 充气式空气床垫打鼾检测和响应
CN107025783A (zh) * 2016-02-02 2017-08-08 麒盛科技股份有限公司 一种智能电动床系统
CN109259952A (zh) * 2018-08-20 2019-01-25 深圳思迈家科技有限公司 睡眠监控系统及基于该睡眠监控系统的睡眠监控方法
CN109998766A (zh) * 2019-04-15 2019-07-12 深圳市云智眠科技有限公司 一种升降式缓解打鼾的方法及系统
CN112603103A (zh) * 2020-12-15 2021-04-06 慕思健康睡眠股份有限公司 全自动调节方法及全自动智能床垫
CN112820319A (zh) * 2020-12-30 2021-05-18 麒盛科技股份有限公司 一种人类鼾声识别方法及其装置
CN113599051A (zh) * 2021-07-15 2021-11-05 麒盛科技股份有限公司 一种基于深度学习算法的电动床控制方法、系统和计算机程序

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4530698B2 (ja) * 2004-03-25 2010-08-25 三洋電機株式会社 いびき抑制方法及び装置
JP2007236839A (ja) * 2006-03-13 2007-09-20 Ame O 耐雑音性いびき音検出装置
US20160007914A1 (en) * 2014-07-14 2016-01-14 Jason Xu Sleep control device
WO2016115766A1 (zh) * 2015-01-24 2016-07-28 深圳市前海安测信息技术有限公司 用于睡眠质量管理的按摩枕及其控制方法
CN104800061A (zh) * 2015-04-30 2015-07-29 深圳市前海安测信息技术有限公司 用于睡眠质量管理的按摩枕及其控制方法
CN105193566B (zh) * 2015-10-09 2018-04-13 东莞市贸天精密五金制品有限公司 一种抑制打鼾的方法及智能床
CN108451503A (zh) * 2018-03-12 2018-08-28 卢纪元 生物反馈疗法机制和基于该机制的智能睡眠诊疗系统
KR102265714B1 (ko) * 2020-01-23 2021-06-16 주식회사 바디프랜드 사용자의 코골이를 방지하는 전동 침대
CN111665754A (zh) * 2020-06-02 2020-09-15 方应龙 一种智能床的控制装置及实现方法
CN112190386A (zh) * 2020-09-11 2021-01-08 深圳数联天下智能科技有限公司 一种止鼾设备的控制方法及相关装置

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2005152537A (ja) * 2003-11-28 2005-06-16 Yokogawa Electric Corp 制御システム
CN105517464A (zh) * 2013-03-14 2016-04-20 择舒公司 充气式空气床垫打鼾检测和响应
CN104188638A (zh) * 2014-09-09 2014-12-10 喜临门家具股份有限公司 一种睡眠保障系统及其控制方法
CN107025783A (zh) * 2016-02-02 2017-08-08 麒盛科技股份有限公司 一种智能电动床系统
CN109259952A (zh) * 2018-08-20 2019-01-25 深圳思迈家科技有限公司 睡眠监控系统及基于该睡眠监控系统的睡眠监控方法
CN109998766A (zh) * 2019-04-15 2019-07-12 深圳市云智眠科技有限公司 一种升降式缓解打鼾的方法及系统
CN112603103A (zh) * 2020-12-15 2021-04-06 慕思健康睡眠股份有限公司 全自动调节方法及全自动智能床垫
CN112820319A (zh) * 2020-12-30 2021-05-18 麒盛科技股份有限公司 一种人类鼾声识别方法及其装置
CN113599051A (zh) * 2021-07-15 2021-11-05 麒盛科技股份有限公司 一种基于深度学习算法的电动床控制方法、系统和计算机程序

Also Published As

Publication number Publication date
CN113599051A (zh) 2021-11-05

Similar Documents

Publication Publication Date Title
WO2023284814A1 (zh) 一种基于深度学习算法的电动床控制方法、系统和计算机程序
KR102403257B1 (ko) 홈 디바이스 제어 장치 및 동작 방법
CN109259952A (zh) 睡眠监控系统及基于该睡眠监控系统的睡眠监控方法
US20150128353A1 (en) Sleep monitoring system and method
CN102526879B (zh) 用于起搏频率自适应调节的装置和系统
CN109393813A (zh) 智能电动床及其控制方法
CN106618040A (zh) 枕头和睡眠干预方法
CN107875496A (zh) 一种智能睡眠管理眼罩装置及其控制方法
CN110604859B (zh) 基于智能家居设备的辅助睡眠控制方法及系统
CN105422492B (zh) 风扇的控制系统及风扇的控制方法
CN204863622U (zh) 用于睡眠质量管理的按摩设备
CN104800061A (zh) 用于睡眠质量管理的按摩枕及其控制方法
CN113181501A (zh) 一种以智能床垫为中心的物联网健康监测方法
US20220117776A1 (en) Motorized bedding system and application software
CN114027667A (zh) 一种在离床状态判定方法、装置、智能床垫及介质
CN113503634A (zh) 用于空调控制的方法、装置及空调
CN107837075A (zh) 梦境监测装置及监测方法
CN111956001A (zh) 智能床垫的控制方法、智能床垫、装置和存储介质
CN113599053A (zh) 一种气囊枕的自适应调节方法、系统和计算机程序
CN112741614A (zh) 打鼾检测及智能止鼾装置
CN112932225B (zh) 智能唤醒枕头以及基于智能唤醒枕头的唤醒方法
CN117442161A (zh) 一种ai自适应智能床垫
CN112162486A (zh) 一种床垫控制系统
WO2023284813A1 (zh) 一种基于深度学习算法的鼾声监测方法与系统以及相应的电动床控制方法和系统
CN116115198A (zh) 一种基于生理体征的低功耗鼾声自动记录方法及装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22841447

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE