WO2024172189A1 - Sleep evaluation method and sleep evaluation apparatus using sleep stage change pattern - Google Patents

Abstract

Disclosed are a sleep evaluation method and sleep evaluation apparatus using a sleep stage change pattern. The sleep evaluation method according to the present invention begins with a step of receiving, as an input, sleep data measured through a sleep care apparatus and accumulatively storing sleep evaluations. Also, from among same, on the basis of sleep data having a sleep evaluation above a preset threshold, a personalized standard sleep hypnogram is generated. In addition, the sleep evaluation method comprises a step of calculating a sleep pattern similarity between a daily sleep hypnogram for the measured sleep data and the personalized standard sleep hypnogram and visually providing the calculated sleep pattern similarity and a sleep index corresponding thereto. Accordingly, the quality of sleep may be evaluated with greater accuracy by performing an evaluation of sleep of a user according to respective changes in clustered customized sleep stages.

Description

Sleep evaluation method and sleep evaluation device using sleep stage change patterns

The present invention relates to a sleep evaluation method and a sleep evaluation device using a sleep stage change pattern, and more specifically, to a sleep evaluation method and a sleep evaluation device for comparing and evaluating sleep quality using a sleep stage change pattern based on sleep data collected using a sleep care device.

In managing one's health, maintaining a good quality sleep state is very important. Therefore, people make a lot of efforts to get a good night's sleep. In this regard, research on methods for collecting biosignals during human sleep to monitor sleep state or inducing a good night's sleep through appropriate stimulation is increasing.

Recently, research is being conducted on technology to analyze sleep data using deep learning model technology.

Specifically, Korean Patent Document Publication No. 10-2023-0013301 (hereinafter, "Prior Document 1") discloses a method for predicting sleep stages using artificial intelligence by inputting raw biosignal data (EEG, EOG, etc.). In addition, U.S. Patent Document Patent No. US 10,098,582 (hereinafter, "Prior Document 2") discloses a technology for characterizing an individual's sleep time pattern information to facilitate appropriate treatment or treatment selection, and classifies an individual's sleep disorder according to the chaos value of the data using the Lyapunov index.

Prior literature 1 and 2 may be useful for predicting a person's sleep stage or distinguishing sleep disorders, but they only provide analysis of the results of sleep. In other words, there are limitations in performing a more customized and accurate sleep evaluation of an individual and confirming it intuitively.

The present invention aims to solve the above-mentioned problems and other problems, and provides a sleep evaluation method and a sleep evaluation device using sleep stage change patterns, which can perform an individual's sleep evaluation more accurately and in a more customized manner according to a personalized sleep stage change pattern.

In addition, the present invention aims to provide a sleep evaluation method and a sleep evaluation device capable of accurately evaluating sleep quality according to changes in sleep stages by using a customized standard sleep curve (Good Sleep Hypnogram) learned by clustering the user's sleep data and applying a hidden Markov model.

In addition, the present invention aims to provide a sleep evaluation method and a sleep evaluation device that provide a visual comparison between a customized standard sleep curve (Good Sleep Hypnogram) and a daily sleep curve (Daily Hypnogram) for sleep evaluation so that the comparison can be intuitively understood at a glance.

In addition, the present invention aims to provide a sleep evaluation method and a sleep evaluation device that enable a user to check his or her own sleep evaluation by period or simultaneously check the results compared with other groups under selected conditions.

In the present invention, sleep data with good sleep evaluation among measured sleep dates are clustered, and a 'personalized standard sleep curve (Good Sleep Hypnogram)' can be generated as a standard for comparing changes in sleep stages based on the clustered sleep data. In addition, the sleep index corresponding to the degree of sleep pattern consistency between the standard sleep curve generated in this way and the measured daily sleep curve is provided in a visualized form so that the user can check it at a glance.

Specifically, a sleep evaluation method using a sleep stage change pattern according to an embodiment of the present invention is performed including the steps described above. The method may include a step of receiving sleep data measured through a sleep care device and accumulatively storing sleep evaluations, and a step of generating a personalized standard sleep curve based on sleep data having a sleep evaluation higher than a preset threshold. In addition, the method may include a step of calculating sleep pattern similarity between a daily sleep curve for measured sleep data and the personalized standard sleep curve; and a step of visually providing the calculated sleep pattern similarity and a sleep index corresponding thereto.

In an embodiment, the sleep index is a sleep quality index corresponding to the calculation result of the sleep pattern similarity, and the sleep quality index can be determined as a size value proportional to the sleep pattern similarity.

In an embodiment, the visually providing step may include a step of simultaneously displaying on one screen each sleep wave and the sleep index corresponding to the daily sleep curve and the personalized standard sleep curve for the measured sleep data.

In an embodiment, if there are multiple daily sleep curves for the measured sleep data, the step of visually providing may include the step of displaying all of the multiple daily sleep curves; and the step of displaying a sleep wave corresponding to a daily sleep curve selected from the multiple daily sleep curves by overlapping it with a sleep wave corresponding to the personalized standard sleep curve.

In an embodiment, the visually providing step may include a step of displaying a sleep index for a daily sleep curve that satisfies a preset condition among the plurality of daily sleep curves.

In an embodiment, the method may include a step of changing a sleep index for a daily sleep curve satisfying the preset condition to a sleep index for the selected daily sleep curve and displaying the change.

In an embodiment, the step of cumulatively storing the sleep evaluation may include a first process of generating a daily sleep curve based on sleep data measured through a sleep care device; a second process of performing a sleep evaluation on the generated daily sleep curve; a third process of storing sleep data and the sleep evaluation corresponding to the generated daily sleep curve; and a fourth process of repeating the first to third processes for a certain period of time.

In an embodiment, the step of generating the personalized standard sleep curve may be a step of generating the personalized standard sleep curve by applying a sleep pattern analysis probability model to sleep data having a sleep evaluation greater than or equal to a preset threshold.

In an embodiment, the step of generating the personalized standard sleep curve may include a step of performing learning on sleep data having a sleep evaluation greater than or equal to the threshold by applying a Hidden Markov Model (HMM) that defines multiple sleep stages as states as the sleep pattern analysis probability model.

In an embodiment, the step of calculating the sleep pattern similarity may be a step of calculating it using a DTW (Dynamic Time Warping) algorithm.

In an embodiment, the step of calculating the sleep pattern similarity may be a step of calculating using a structure-based similarity calculation algorithm that is performed by calculating the Fisher-Lao distance and the Gini coefficient considering the similarity probability between states included in the sleep pattern analysis probability model.

In an embodiment, the step of visually providing may include the step of providing a user interface for displaying the calculated sleep pattern similarity and the corresponding sleep index on a daily, weekly, or monthly basis.

In an embodiment, the method may include the steps of: receiving a sleep analysis request prior to the step of generating the personalized standard sleep curve; determining, in response to the sleep analysis request, that the personalized standard sleep curve is not to be generated; and displaying visual information indicating that sleep data is being collected to provide the sleep index in response to the determination.

In an embodiment, the step of visually providing may further include the step of providing a comparison of the sleep index and one or more related sleep assessment factors with a first group of sleep indices corresponding to selected filter items.

In an embodiment, the method may include, in the step of providing the comparison, a step of receiving an input for changing the filter item; a step of updating the sleep index based on the filter item changed according to the input; and a step of comparing the updated sleep index and one or more sleep assessment factors related thereto with a sleep index of a second group corresponding to the filter item changed according to the input.

In addition, in an embodiment of the present invention, a sleep evaluation device using a sleep stage change pattern may include a communication unit which receives sleep data of a user measured through a sleep care device; a data processing unit which accumulates and stores sleep evaluations for the measured sleep data in a storage device and generates a personalized standard sleep curve based on sleep data having a sleep evaluation higher than a preset threshold; and a display unit. Here, the data processing unit may calculate a sleep pattern similarity between a daily sleep curve for the measured sleep data and the generated personalized standard sleep curve, and control the display unit to visually display the calculated sleep pattern similarity and a sleep index corresponding thereto.

In an embodiment, the sleep index is a sleep quality index corresponding to the calculation result of the sleep pattern similarity, and the sleep quality index can be determined as a size value proportional to the sleep pattern similarity.

In an embodiment, the display unit can simultaneously display each sleep wave and the sleep index corresponding to the daily sleep curve and the personalized standard sleep curve for the measured sleep data on one screen.

In an embodiment, the data processing unit may apply a sleep pattern analysis probability model to sleep data having a sleep evaluation greater than or equal to the preset threshold value to generate the personalized standard sleep curve.

In an embodiment, the data processing unit may perform learning on sleep data having a sleep evaluation greater than or equal to the threshold by applying a Hidden Markov Model (HMM) that defines multiple sleep stages as states as the sleep pattern analysis probability model.

According to an embodiment of the present invention, the following effects can be achieved.

According to at least one of the embodiments of the present invention, by clustering sleep data evaluated as having good sleep quality among the user's sleep data and utilizing a customized and personalized standard sleep curve (Good Sleep Hypnogram) as a comparison standard based on the clustered sleep data, the user's sleep evaluation is performed for each clustered customized sleep stage change, thereby enabling a more accurate evaluation of sleep quality.

In addition, according to at least one of the embodiments of the present invention, it is possible to intuitively understand at a glance the change in sleep stages for the daily sleep curve (Daily Hypnogram) to be compared with the customized standard sleep curve (Good Sleep Hypnogram).

In addition, according to at least one of the embodiments of the present invention, a user can easily check his/her own sleep evaluation for a desired period of time and simultaneously check the results compared to other groups under selected conditions.

FIG. 1 is a drawing illustrating a sleep care device according to an embodiment of the present invention and devices operating in conjunction therewith.

Figure 2 is a block diagram showing the detailed configuration of the sleep care device and terminal of Figure 1.

FIGS. 3 and 4 are a front perspective view and a back perspective view of a sleep care device according to an embodiment of the present invention.

Figure 5 illustrates the sleep care device of Figure 3 being worn on the user's ear.

FIG. 6a is an exploded perspective view of the sleep care device of FIG. 3, and FIG. 6b illustrates components (hardware) of the sleep care device mounted on or electrically connected to the substrate of FIG. 6a.

Figures 7 to 9 illustrate a plan view, a bottom view, and an enlarged view of A of Figure 8 of the ear hook of the sleep care device of Figure 3.

FIG. 10 is a drawing illustrating a process of a sleep evaluation method using a sleep care device according to an embodiment of the present invention.

FIGS. 11A and 11B are example screens visually displaying the similarity of sleep patterns and the corresponding sleep index according to changes in the user's sleep stages, according to an embodiment of the present invention.

FIG. 12 is an example screen related to a data collection method for generating a standard sleep curve according to an embodiment of the present invention.

FIG. 13 is an exemplary flowchart illustrating a method for calculating a slip index according to an embodiment of the present invention.

Figure 14a shows a human sleep cycle, and Figure 14b is an example showing a human brain wave measured by a sleep care device.

FIG. 15 illustrates a hidden Markov model as a machine learning model for generating a standard sleep curve according to an embodiment of the present invention, and FIG. 16 is an example visualizing a standard sleep curve learned using the model of FIG. 15.

FIG. 17a is an example screen for comparing records of a user's sleep index according to an embodiment of the present invention, and FIG. 17b is an example screen displayed when requesting a sleep index before generating a personalized standard sleep curve.

FIGS. 18a and 18b are example screens that visually provide the results of comparing the sleep index of a group corresponding to a filtering condition and the sleep index of a user according to an embodiment of the present invention.

Hereinafter, embodiments disclosed in this specification will be described in detail with reference to the attached drawings. Regardless of the drawing symbols, identical or similar components will be given the same reference numerals and redundant descriptions thereof will be omitted. The suffixes "module" and "part" used for components in the following description are assigned or used interchangeably only for the convenience of writing the specification, and do not have distinct meanings or roles in themselves. In addition, when describing the embodiments disclosed in this specification, if it is determined that a specific description of a related known technology may obscure the gist of the embodiments disclosed in this specification, the detailed description thereof will be omitted. In addition, the attached drawings are only intended to facilitate easy understanding of the embodiments disclosed in this specification, and the technical ideas disclosed in this specification are not limited by the attached drawings, and should be understood to include all modifications, equivalents, and substitutes included in the spirit and technical scope of the present invention.

Hereinafter, a sleep induction device according to the present disclosure and an operating method thereof will be specifically described. The sleep induction device according to the present disclosure is configured to be wearable on a part of a user's body and includes a sensor having a structure that can be coupled and separated.

The 'sleep care device' disclosed in this specification is formed as a wearable structure on a part of the user's body, and can sense the user's biosignals while the user is in a sleeping state and perform wireless communication (or data communication) by connecting to at least one external device.

The 'terminal', 'device', and 'external device' disclosed in this specification are connected to a 'sleep care device' so as to be able to communicate with it, receive the user's sleep data measured through the sleep care device, and display the sleep evaluation results on the screen.

The 'sleep evaluation device' disclosed in the present invention is directly or indirectly connected to a 'sleep care device', receives sleep data of a user measured through the sleep care device, evaluates sleep, accumulates the data to create a personalized standard sleep curve (good sleep hypnogram), and calculates the similarity of sleep patterns by sleep stage and daily curve corresponding to the measured sleep data to provide a corresponding sleep index.

In addition, the operation executed in the above-described 'sleep evaluation device' may also be performed by executing a specific application in the above-described 'terminal', 'device', or 'external device' and connecting to an external server/cloud/device/system. In this case, the operation executed in the above-described 'sleep evaluation device' may be implemented identically or similarly through the execution of an application in the 'terminal', 'device', or 'external device' and thereby, and it is to be noted in advance that this may be equally applied throughout the specification described below.

Previously, when evaluating sleep quality, sleep scores were calculated mainly by considering time values such as the total time the user slept and the sum of the time for each sleep stage.

However, in the present invention, in evaluating the quality of sleep of a user, the quality of sleep is implemented so that it can be evaluated more accurately according to the change in sleep stage of a learned personalized standard sleep curve (Good Sleep Hypnogram).

FIG. 1 illustrates a sleep care device (100) according to an embodiment of the present invention, devices operating in conjunction therewith, such as a terminal (200) and a sleep evaluation device (300).

Referring to FIG. 1, a sleep care device (100) according to an embodiment of the present invention is mounted on a part of a user's body and measures sleep data such as biosignals of the user while sleeping, for example, brain wave signals.

The measured sleep data is transmitted to the terminal (200), and when the measurement is completed, a daily sleep curve (daily hypnogram) for the measured sleep data is created and a sleep evaluation is performed.

The above daily hypnogram is an index/graph representation of sleep stage-specific changes (e.g., wake, REM, Non-REM 1/2/3/4 stages) based on measured sleep data. In addition, the sleep evaluation is a numerical representation of sleep quality for the measured sleep data, and is performed after the sleep data measurement is completed. The sleep evaluation may be determined as one of the values within the range of, for example, score 1 to score 5.

The sleep evaluation device (300) communicates with the terminal (200) (and/or the sleep care device (100)) via the network (50) to receive measured sleep data and a corresponding sleep evaluation. The sleep evaluation device (300) can cumulatively store the received measured sleep data and sleep evaluation in a storage device (350).

The storage device (350) may be an EMR (Electronic Medical Record) device for storing sleep data. In addition, the storage device (350) may store a machine learning model, i.e., a sleep pattern analysis probability model, for generating a personalized standard sleep curve (Good Sleep Hypnogram) generated by the sleep evaluation device (300).

The sleep evaluation device (300) may be implemented by including an external device (e.g., external device/server/cloud/system) or application (310) to perform a process of generating a personalized standard sleep curve (Good Sleep Hypnogram), calculating the similarity with the daily sleep curve (daily hypnogram), and outputting a sleep index corresponding to the calculated result.

In this case, the external device or application (310) can operate in conjunction with the storage device (350). The external device or application (310) can receive measured sleep data and sleep evaluation through the data input device (311). The external device or application (310) can apply a sleep pattern analysis probability model to sleep data having a sleep evaluation higher than a preset threshold through the data analysis device (312) to generate a personalized standard sleep curve.

Here, the reason why a sleep evaluation above a preset threshold is required is because clustered sleep data evaluated as having high sleep quality are required to generate a personalized standard sleep curve (Good Sleep Hypnogram). The preset threshold can be, for example, a score 4 or score 5 value when the sleep evaluation range is score 1 to score 5.

In addition, an external device or application (310) can calculate a sleep pattern similarity between the generated personalized standard sleep curve and the daily sleep curve through a data analysis device (312), and provide a sleep index (sleep) corresponding to the calculated sleep pattern similarity. At this time, the sleep index (sleep) is a sleep quality index corresponding to the calculation result of the sleep pattern similarity, and the size of the sleep quality index is determined in proportion to the degree of sleep pattern similarity.

An external device or application (310) can call a sleep pattern analysis probability model and an algorithm for calculating sleep pattern similarity from a data storage device (313), and store data on a personalized standard sleep curve (Good Sleep Hypnogram) in the data storage device (313).

An external device or application (310) can simultaneously display the generated personalized standard sleep curve and the sleep index corresponding to the sleep wave and sleep pattern similarity of each of the daily sleep curves through the data output device (314). Alternatively, the external device or application (310) can transmit related data/information through the network (50) so that visualization of each sleep wave and sleep index can be performed through the display unit of the terminal (200).

Meanwhile, as described above, the operation executed in the sleep evaluation device (300), that is, the operation implemented in an external device (e.g., external device/server/cloud/system) or application (310), may also be executed through a specific application installed in the terminal (200).

Although not directly illustrated in FIG. 1, the sleep evaluation device (300) may be implemented to include a communication unit for communicating directly or indirectly with the sleep care device (100) and/or the terminal (200), a data processing unit (or a control unit/processor) for performing a series of processes for generating a personalized standard sleep curve, calculating a sleep pattern similarity, and generating a corresponding sleep index, and a display unit (e.g., a touch screen) for visually displaying the calculated sleep pattern similarity and the corresponding sleep index.

In this case, the sleep evaluation device (300) can perform operations including those performed by the terminal (200). In addition, in this case, the display unit of the sleep evaluation device (300) can simultaneously display, on one screen, each sleep wave and sleep index corresponding to the daily sleep curve for the measured sleep data and the personalized standard sleep curve. In addition, in this case, the data processing unit of the sleep evaluation device (300) can generate the personalized standard sleep curve by applying and learning a Hidden Markov Model (HMM) that defines multiple sleep stages as states for sleep data having a sleep evaluation higher than a threshold value.

FIG. 2 is a block diagram showing the detailed configuration of the sleep care device (100) of FIG. 1 and an external device (200) linked thereto, such as a terminal.

Below, FIG. 2 is a block diagram for explaining each configuration of a sleep care device (100) according to the present disclosure and an external device (200) communicating therewith. The sleep care device (100) and the external device (200) illustrated in FIG. 2 may include only some of the illustrated configurations or may additionally include other configurations.

The sleep care device (100) may include a communication unit (101), a signal processing unit (102), a memory unit (103), a sensing unit (104), an output unit (105), a power supply unit (109), and a control unit (108). The communication unit (101) or the communication module may communicate with an external device (200) using a wireless communication network such as WiFi, LiFi, etc., or a short-range communication network such as Bluetooth, RFID (Radio Frequency Identification), infrared communication (IrDA, infrared Data Association), UWB (Ultra Wideband), ZigBee, etc.

The signal processing unit (102) can preprocess (e.g., noise removal, offset correction, calculation of RMS/PSD, etc.) and process the signals acquired by the sensing unit (104) for analysis, and provide them to the communication unit (101) and/or the memory unit (103).

The sensing unit (104) may include a plurality of sensors for obtaining a sleep-related signal. Specifically, the sensing unit (104) may be detachably coupled to the main body of the sleep care device (100) and may include one or more sensors that contact a part of the user's body to obtain a sleep-related signal. In addition, the sensing unit (104) may further include a touch sensor, a pressure sensor, etc. for recognizing an input related to the operation of the sleep care device (100).

The one or more sensors may include, for example, sensors for obtaining bio-signals, such as an EEG sensor, a PPG sensor, an EDA sensor, an ECG sensor, an EMG sensor, and the like; sensors for obtaining signals related to user movement, such as an acceleration sensor, a gyro sensor, and a motion sensor; and/or sensors for obtaining signals related to the environment of a sleeping space, such as a light sensor, a temperature sensor, a humidity sensor, and a noise measurement sensor. In one embodiment, the sensor for obtaining the user's bio-signals (e.g., brain wave signals) may include a plurality of electrode sensors (at least two or more electrode sensors). For example, when the sensor for obtaining the user's bio-signals includes three electrode sensors, each of them may be a measurement electrode (SIG), a reference electrode (REF), and a ground electrode (GND) for measuring EEG signals. In one embodiment, at least one of the plurality of electrode sensors may be placed to contact the user's in-ear.

In addition, some of the plurality of electrode sensors, for example, the measuring electrode (SIG) and the reference electrode (REF), may be positioned to be spaced apart from each other in order to obtain a strong signal strength (Voltage (RMS)). For example, the measuring electrode (SIG) may be positioned to contact the in-ear, and the reference electrode (REF) may be positioned to contact the back of the user's ear. More preferably, in order to secure a further separation distance, the measuring electrode (SIG) may be implemented to contact the in-ear of the left (right) ear, and the reference electrode (REF) and the ground electrode (GND) may be implemented to contact the right (left) ear.

The memory unit (103) can store sleep-related signals acquired through the sensing unit (104), such as EEG signals, movement signals, etc. When storing sleep-related signals, the memory unit (103) can store metadata such as the reception time of the corresponding signal and the most recent sleep state (or sleep stage). In addition, the memory unit (103) can store the user's sleep pattern (e.g., time required to fall asleep, REM entry point, etc.) and the output pattern of the synthesized sound source. In addition, the memory unit (103) can store the notification time and information related thereto.

The memory unit (103) may include at least one type of storage medium among a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (for example, an SD or XD memory, etc.), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, and an optical disk.

The output unit (105) or output module outputs a synthetic sound source for inducing the user to sleep. To this end, the output unit (205) may include a receiver, a speaker, a buzzer, etc. When the sleep care device (100) is respectively mounted on each of the user's ears, each output unit (105) of each sleep care device (100) may output a synthetic sound source having a different frequency band. In another embodiment, when the sleep care device (100) is implemented as a glasses-type device that comes into contact with the user's ears, the output unit (105) may include a display that can output visual information.

The control unit (108) controls the overall operation of the sleep care device (100). The control unit (208) can transmit a sleep-related signal obtained through the sensing unit (104) to an external device (200) through the communication unit (101), and can receive information on the user's sleep status based on analysis of the sleep-related signal from the external device (200).

The power supply unit (109) supplies power to the sleep care device (100) and may include a rechargeable battery module.

The external device (200) communicates with the sleep care device (100) and may include a communication unit (221), a sensing unit (224), a display unit (225), and a data processing unit (228). The external device (200) may be any computing device capable of wired and/or wireless communication. For example, the external device (200) may include a smart phone, a mobile phone, a computer, a laptop, a PDA (Personal Digital Assistants), a PMP (Portable Multimedia Player), a tablet PC, other wearable devices, etc. In addition, the external device (200) may include a preset application for interacting with the sleep care device (100) or may connect to a server to download a preset application.

The sensing unit (224) of the external device (200) may include information related to the environment of the sleeping place, such as a temperature sensor, a humidity sensor, a light sensor, a noise detection sensor, etc. The data processing unit (228) of the external device (200) may analyze a sleep-related signal obtained from the sleep care device (100), determine the user's sleep state based on the analysis result, and transmit information about the sleep state to the sleep care device (100). The display unit (225) of the external device (200) may visually display screen information indicating the sleep state, and may display sleep pattern history information in response to a user's inquiry request.

In one embodiment, in response to detecting that the sleep care device (100) is worn on a part of the user's body, the sleep care device (100) and the external device (200) may be connected to communicate. To this end, the sleep care device (100) may be pre-registered with the external device (200), or the external device (200) may be pre-registered with the sleep care device (100). Alternatively, the sleep care device (100) and the external device (200) may be connected according to the execution of a preset application installed in the external device (200). Additionally, in one embodiment, in response to detecting that the sleep care device (100) is separated from the user's body, the sleep care device (100) and the external device (200) may be disconnected.

The external device (200) communicates with the sleep evaluation device (300) via the network (50) as described in FIG. 1 and transmits sleep data acquired from the sleep care device (100). In addition, operations according to the embodiment of the present invention executed by the sleep evaluation device (300) may also be performed via the external device (200) depending on the execution of a specific application installed in the external device (200). In this case, collection of measured sleep data, sleep evaluation, calculation of personalized standard sleep curves, sleep pattern similarity, generation of a corresponding sleep index, and provision of visualization may all be performed using the external device (200).

Hereinafter, the sleep care device (100) used in the present invention will be specifically described with reference to FIGS. 3 to 9. However, the sleep care device (100) described below may be implemented to include more or fewer components than those illustrated, as needed.

As shown in FIGS. 3 to 9, a sleep care device (100) according to an example of an embodiment of the present invention includes an ear hook (10), an ear tip (20), and a housing (30).

Specifically, FIGS. 3 and 4 are a front perspective view and a back perspective view of a sleep care device (100) according to an embodiment of the present invention. FIG. 5 illustrates a state in which the sleep care device (100) of FIG. 3 is worn on a user's ear. FIG. 6a is an exploded perspective view of the sleep care device (100) of FIG. 3, and FIG. 6b illustrates components (hardware) of the sleep care device (100) mounted on or electrically connected to the substrate of FIG. 6a. FIGS. 7 to 9 illustrate a plan view, a bottom view, and an enlarged view of A of FIG. 8 of an ear hook (10) of the sleep care device (100) of FIG. 3.

The ear hook (10) serves to secure the sleep care device (100) of the present invention to the ear. Due to the ear hook (10), the sleep care device (100) of the present invention is not easily separated from the user's ear.

The ear hook (10) is formed in the shape of a string (or line) having a preset length.

Specifically, the ear hook (10) includes one end (111) and the other end (122) opposite to the one end (111), and is formed in a loop (open loop) shape having a preset length extending from the one end (111) to the other end (122).

In addition, the ear hook (10) has a preset cross-sectional diameter. Here, the cross-section refers to a cross-section when the ear hook (10) is cut in a direction orthogonal to the longitudinal direction of the ear hook (10). The shape of the cross-section may be circular, oval, or polygonal.

The ear hook (10) is connected to the housing (30). The ear hook (10) is connected to the housing (30) to form a closed loop. An auricle is inserted and placed into an opening (10h) within the closed loop.

Specifically, one end (111) of the ear hook (10) is connected to one side (311) of the housing (30), and the other end (122) of the ear hook (10) is connected to the other side (312) of the housing (30). Here, the other side (312) of the housing (30) represents a part of the housing (30) adjacent to the one side (311) of the housing (30) and spaced apart from the one side (311) by a preset distance. The one side (311) of the housing (30) represents one side (311) of a lower housing (31) described below, and the other side (312) of the housing (30) represents the other side (312) of the lower housing (31).

When the sleep care device (100) is mounted on the user's ear, the ear tip (20) described below is placed at the entrance of the external auditory canal (concha) of the ear. Then, the housing (30) connected to the ear tip (20) is placed so as to face the ear with the ear tip (20) in between. Then, the ear hook (10) connected to the housing (30) forms a closed loop together with the housing (30). The auricle is inserted and placed into the opening (10h) in the closed loop, and the ear hook (10) is placed between the head and the auricle to wrap around the auricle. Therefore, the ear hook (10) serves to fix the sleep care device (100) of the present invention so that it is not easily separated from the user's ear.

The ear hook (10) is formed of a material (or material) having flexibility and elasticity. For example, the ear hook (10) may be formed of a silicone material having flexibility and elasticity. Here, flexibility refers to the property of being stretchable (elastic), bendable, and flexible. Elasticity refers to the property of an object (ear hook (10)) that has been deformed by an external force to return to its original state when the external force is removed.

The ear hook (10) is formed of a flexible and elastic material, and is placed between the head and the auricle to maintain contact (close contact) with the head and/or the auricle.

Meanwhile, the ear hook (10) includes a first electrode (11), a second electrode (12), and a joint (13).

The first electrode (11) includes one end (111) and another end (112) opposite to the one end (111). The first electrode (11) is formed with a preset length extending from the one end (111) to the other end (112).

The above one end (111) of the first electrode (11) represents one end (111) of the ear hook (10) and is connected to one side (311) of the housing (30). The other end (112) of the first electrode (11) is connected to the connecting portion (13) (specifically, one end (131) of the connecting portion (13)).

Since the first electrode (11) is formed to a preset length, even if a part of the first electrode (11) does not come into contact with the head and/or the auricle, another part of the first electrode (11) can come into contact with the head and/or the auricle. Accordingly, when the sleep care device (100) of the present invention is mounted on the ear, the first electrode (11) always remains in contact with the head and/or the auricle.

The first electrode (11) is formed in an arc shape to surround the auricle. Here, the auricle may represent, for example, the auricle corresponding to the upper part of the ear.

One end (111) of the ear hook (10), that is, one end (111) of the first electrode (11), is connected to one side (311) of the housing (30). Specifically, one end (111) of the first electrode (11) is coupled to one side (311) of the lower housing (31) described below.

One end (111) of the first electrode (11) can be formed to correspond to the shape of the portion to be joined in the lower housing (31). However, one end (111) of the first electrode (11) is not limited to a specific shape and can be formed in various shapes.

One end (111) of the first electrode (11) is formed with a plurality of fastening holes. The plurality of fastening holes include one or more fastening projection holes (1111) and fastening screw holes (1112). Two fastening projection holes (1111) may be formed, and one fastening screw hole (1112) may be formed.

The above fastening projection hole (1111) is positioned so that a fastening projection formed on one side (311) of the lower housing (31) is inserted. As a result, one end (111) of the ear hook (10) is fixed to the lower housing (31). The above fastening screw hole (1112) is seated in a fastening screw groove formed on one side (311) of the lower housing (31).

A fastening screw (311S) passes through a fastening screw hole formed in the substrate (34) and a fastening screw hole (1112) formed in one end (111) of the first electrode (11), and is inserted into and fixed in a fastening screw groove (3112) formed in one side (311) of the lower housing (31). As a result, the substrate (34) and one end (111) of the first electrode (11) are fixed to one side (311) of the lower housing (31).

A plurality of support protrusions (1113) can be formed around the fastening screw hole (1112) at one end (111) of the first electrode (11).

The above-mentioned plurality of support protrusions (1113) protrude from one end (111) of the first electrode (11) toward the substrate (34) and are arranged spaced apart from each other at a preset interval.

A plurality of support protrusions (1113) support the substrate (34) and space one end (111) of the first electrode (11) from the substrate (34). Accordingly, one end (111) of the first electrode (11) does not directly contact the substrate (34), but comes into contact with a specific electric circuit formed on the substrate (34) through a metal clip (341) described later.

The second electrode (12) includes one end (121) and the other end (122) opposite to the one end (121). The second electrode (12) is formed with a preset length extending from the one end (121) to the other end (122).

The above-mentioned one end (121) of the second electrode (12) is connected to the joint (13) (specifically, the other end (132) of the joint (13)). The above-mentioned other end (122) of the second electrode (12) represents the other end (122) of the ear hook (10) and is connected to the other side (312) of the housing (30).

Since the second electrode (12) is formed with a preset length, even if some part of the second electrode (12) does not come into contact with the head and/or the auricle (or earlobe), some other part of the second electrode (12) can come into contact with the head and/or the auricle (or earlobe). Accordingly, when the sleep care device (100) of the present invention is mounted on the ear, the second electrode (12) always remains in contact with the head and/or the auricle (or earlobe).

The second electrode (12) is formed in an arc shape to surround the auricle (or earlobe). Here, the auricle (or earlobe) may refer to, for example, the auricle (or earlobe) corresponding to the lower part of the ear.

The other end (122) of the ear hook (10), that is, the other end (122) of the second electrode (12), is connected to the other side (312) of the housing (30). Specifically, the other end (122) of the second electrode (12) is coupled to the other side (312) of the lower housing (31) described below.

The other end (122) of the second electrode (12) can be formed to correspond to the shape of the portion to be joined in the lower housing (31). However, the other end (122) of the second electrode (12) is not limited to a specific shape and can be formed in various shapes.

The other end (122) of the second electrode (12) is formed with a plurality of fastening holes.

The plurality of fastening holes formed in the other end (122) of the second electrode (12) include one or more fastening projection holes (1221) and fastening screw holes (1222). Two fastening projection holes (1221) may be formed, and one fastening screw hole (1222) may be formed.

The fastening projection hole (1221) formed on the other end (122) of the second electrode (12) is positioned so that the fastening projection formed on the other side (312) of the lower housing (31) is inserted therein. As a result, it is fixed to the other end (122) of the ear hook (10).

The fastening screw hole (1222) formed in the other end (122) of the second electrode (12) is seated in the fastening screw groove formed in the other side (312) of the lower housing (31).

The fastening screw (312S) passes through the fastening screw hole formed in the substrate (34) and the fastening screw hole (1222) formed in the other end (122) of the second electrode (12) and is inserted into and fixed in the fastening screw groove (3122) formed in the other side (312) of the lower housing (31). As a result, the substrate (34) and the other end (122) of the second electrode (12) are fixed to the other side (312) of the lower housing (31).

A plurality of support protrusions (1223) may be formed around the fastening screw hole (1222) at the other end (122) of the second electrode (12).

The plurality of support protrusions (1223) formed on the other end (122) of the second electrode (12) protrude from the other end (122) of the second electrode (12) toward the substrate (34) and are arranged spaced apart from each other at a preset interval.

A plurality of support protrusions (1223) formed on the other end (122) of the second electrode (12) support the substrate (34) and space the other end (122) of the second electrode (12) from the substrate (34). Accordingly, the other end (122) of the second electrode (12) does not directly contact the substrate (34), but rather comes into contact with a specific electric circuit formed on the substrate (34) through a metal clip (341) described later.

The first electrode (11) and the second electrode (12) are electrically conductive. The first electrode (11) and the second electrode (12) are formed of a flexible and elastic material. For example, the first electrode (11) and the second electrode (12) are formed of a flexible and elastic silicon material, and can be formed by including a metal powder having electrical conductivity in the silicon material.

Specifically, the first electrode (11) and the second electrode (12) can be formed by mixing plate-shaped silver (Ag) powder, spherical silver (Ag) powder, and/or silver flakes into a silicone polymer. As a result, the first electrode (11) and the second electrode (12) have electrical conductivity while maintaining the characteristics of the silicone material. The first electrode (11) and the second electrode (12) each have a resistance value of 10Ω (ohm) or less.

The first electrode (11) and the second electrode (12) can sense and measure brain wave signals. Here, brain waves (Electroencephalogram; EEG) are electrical signals generated by neural activity in the brain. Brain waves are also called electroencephalograms.

The first electrode (11) and the second electrode (12) serve as electrodes. For example, the first electrode (11) may be a reference electrode, and the second electrode (12) may be an earth electrode (bias). In addition, the ear tip (20) described below may be a measuring electrode.

The first electrode (11) and the second electrode (12) are formed of a material having flexibility and elasticity, so that they are placed between the head and the auricle and maintain a state of contact (close contact) with the head and/or the auricle. In addition, the first electrode (11) and the second electrode (12) include a metal powder having electrical conductivity, so that they can function as electrodes.

One end (111) of the first electrode (11) is in contact with an electrically conductive metal clip (341) mounted on a substrate (34). The first metal clip (341a) is mounted (fixed) to an electric circuit of the substrate (34). The first metal clip (341a) is arranged between one end (111) of the first electrode (11) and the substrate (34), and electrically connects one end (111) of the first electrode (11) and the electric circuit of the substrate (34). Specifically, one end (e.g., an upper end) of the first metal clip (341a) is in contact with one end (111) of the first electrode (11), and the other end (e.g., a lower end) of the first metal clip (341a) is in contact with the electric circuit of the substrate (34).

Accordingly, the electric signal (specifically, brain waves) measured by the first electrode (11) is transmitted to the substrate (34) via one end (111) of the first electrode (11) and the first metal clip (341a), and is stored in a memory placed on the substrate (34) or processed by a signal processing unit into data for output or communication.

Additionally, the other end (122) of the second electrode (12) comes into contact with an electrically conductive metal clip (341) mounted on the substrate (34).

The above second metal clip (341b) is mounted (fixed) to the electric circuit of the substrate (34).

The second metal clip (341b) is placed between the other end (122) of the second electrode (12) and the substrate (34), electrically connecting the electric circuit of the other end (122) of the second electrode (12) and the substrate (34).

Specifically, one side (e.g., upper side) of the second metal clip (341b) contacts the other end (122) of the second electrode (12), and the other side (e.g., lower side) of the second metal clip (341) contacts the electric circuit of the substrate (34).

Accordingly, the electric signal (specifically, brain waves) measured by the second electrode (12) is transmitted to the substrate (34) via the other end (122) of the second electrode (12) and the second metal clip (341b), and is stored in a memory placed on the substrate (34) or processed by a signal processing unit into data for output or communication.

Instead of the first metal clip (341a) and the second metal clip (341b), an electrically conductive metal terminal, for example, a metal brick or a metal pogo pin, may be applied.

The joint (13) is placed between the first electrode (11) and the second electrode (12) and is formed to a preset length.

The joint (13) insulates the first electrode (11) and the second electrode (12) from each other. The joint (13) is formed of a material having flexibility and elasticity, but does not contain a metal powder having electrical conductivity. Accordingly, the joint (13) is not electrically conductive.

Since the first electrode (11) and the second electrode (12) are separated (spaced) by the joint (13), the first electrode (11) and the second electrode (12) can obtain a clean (reliable) electric signal that does not interfere with each other.

The joint (13) includes one end (131) and another end (132) opposite to the one end (131). The joint (13) is formed with a preset length extending from the one end (131) to the other end (132).

The above one end (131) of the joint (13) is connected to the other end (112) of the first electrode (11). The above other end (132) of the joint (13) is connected to one end (121) of the second electrode (12).

The other end (112) of the first electrode (11) and one end (131) of the joint (13) can be formed integrally. This can be achieved by a double injection method. In addition, the other end (132) of the joint (13) and one end (121) of the second electrode (12) can also be formed integrally as described above.

Alternatively, the other end (112) of the first electrode (11) and one end (131) of the joint (13) may be joined (connected) by a method such as hetero-silicon bonding. The other end (132) of the joint (13) and one end (121) of the second electrode (12) may also be joined (connected) by a method such as hetero-silicon bonding.

The joint (13) is formed in a wave shape (133). In other words, the joint (13) is formed in a winding (bent here and there) shape (133). Or, the joint (13) is formed in a shape (133) that is folded multiple times.

Various electronic components are mounted on the substrate (34) of Fig. 6a. The substrate (34) represents a printed circuit board.

Referring to FIG. 6b, various electronic components may be mounted on the substrate (34), including: a memory for storing brainwave signals measured from the eartip (20); a signal processing unit for converting analog signals into digital signals (digital); an antenna for transmitting data to or receiving data from a mobile terminal; a speaker for outputting sound; a display unit for displaying various information in the case where the sleep care device (1) of the present invention has a display unit; an output unit for outputting sound so that sound is transmitted through the speaker; sensors such as an acceleration sensor, a brainwave sensor including a brainwave electrode, a battery; and electronic components related to controlling signals and data so that the aforementioned functions are organically performed (control unit).

When the sleep care device (100) is worn on the ear, the joint (13) can be easily worn between the ear and the head because the shape (133) (wave shape, winding shape or multiple folded shape) of the joint (13) is spread out. In addition, after being worn, the shape (133) of the joint (13) returns to its original state, so that the ear hook (10) can be in contact (closely attached) to the head and/or the ear as a whole.

In addition, due to the shape (133) of the joint (13) (wave shape, winding shape or shape folded multiple times) and the flexibility and elasticity of the shape (133), there is no need to replace the ear hook (10) depending on the user (specifically, the size of the auricle), and it can be used by multiple family members and other people.

Below, a sleep evaluation method using a sleep stage change pattern according to an embodiment of the present invention based on the user's sleep data input through the aforementioned sleep care device (100) will be specifically examined.

FIG. 10 illustrates a process of a sleep evaluation method using a sleep stage change pattern of sleep data measured by a sleep care device (100) according to an embodiment of the present invention.

Meanwhile, in Fig. 10, the sleep care device (100), the terminal (200), and the sleep evaluation device (300) are shown to operate in conjunction with each other, but this is not limited thereto. For example, it may be performed by the sleep care device (100) and the sleep evaluation device (300) equipped with a display unit, or it may be performed by the terminal (200) and the sleep care device (100) interacting to execute a specific application and perform an operation according to an embodiment of the present invention.

Referring to FIG. 10, when sleep data is measured through a sleep care device (100) while the user is sleeping (S10), the measured sleep data is collected (S20) and a daily sleep curve (daily hypnogram) is generated therefor.

The above daily hypnogram refers to an index/graph that represents the change in sleep stages classified into the user's awake stage (WAKE), light sleep stage (STAGE 1~STAGE 2), deep sleep stage (STAGE 3~STAGE 4), and REM sleep stage.

The daily hypnogram can be generated in real time based on sleep data measured during the user's sleep, or can be generated all at once after the sleep data measurement is completed. In addition, the user can sleep multiple times during the day, in which case a daily sleep curve for each sleep, i.e. multiple daily sleep curves, can be generated.

The collected measured sleep data is transmitted to the sleep evaluation device (300) along with the daily sleep curve (daily hypnogram) and its evaluation score.

The evaluation score for the collected measured sleep data means an evaluation of sleep satisfaction, and for example, one of scores 1 to 5 is given depending on the sleep satisfaction level. The evaluation score for the collected measured sleep data is performed after data measurement using the sleep care device (100) is completed.

The sleep evaluation device (300) accumulates collected measured sleep data and an evaluation score therefor, i.e., data regarding sleep evaluation (S30). For example, measured sleep data collected over a specified period (e.g., 30 days) and an evaluation score performed thereon may be sequentially stored in the sleep evaluation device (300).

In an embodiment, data regarding sleep evaluation may be cumulatively stored in a separate storage device (e.g., a first database). For example, as illustrated in FIG. 1, the sleep evaluation device (300) may receive measured sleep data from the sleep care device (100) via the terminal (200), perform a sleep evaluation on the measured sleep data, and accumulate and store the data in a linked server/database (350, FIG. 1). At this time, the storage device (350) such as the linked server/database may be implemented as being included in the sleep evaluation device (300) (e.g., in the form of a memory) or may be implemented as an external, separate storage device.

After this, filtering (S40) is performed so that only sleep data that satisfies a preset threshold or higher in evaluation scores are considered valid data. For example, if the evaluation score range is score 1 to score 5, score 4 or score 5 can be the threshold, and sleep data having such evaluation scores are filtered as valid data.

The data filtered as the above valid data is stored in a separate storage device (e.g., a second database) and becomes input data for generating a personalized standard sleep curve (Good Sleep Hypnogram). In this way, by clustering sleep data with high evaluation scores as input data, a personalized standard sleep curve (Good Sleep Hypnogram) can be generated with sleep data with good sleep quality.

Next, the process (S50) for generating a personalized standard sleep curve is as follows.

Specifically, the above process (S50) receives input data for generating a personalized standard sleep curve. That is, accumulated sleep data having a sleep evaluation score higher than a threshold are input as input data. Then, sleep data having a sleep evaluation score higher than a threshold are clustered and a personalized standard sleep curve is generated based on this. This clusters sleep data having good sleep quality and generates a standard pattern of sleep stage changes for this, thereby enabling a more customized and accurate sleep quality evaluation for each user.

Specifically, the above process (S50) can perform machine learning using a sleep pattern analysis probability model (ML Algorithm) on the input data.

For example, a Hidden Markov Model (HMM) that defines multiple sleep stages as states is applied to learn about the input data, and a personalized standard sleep curve (Good Sleep hypnogram) is generated accordingly.

Thereafter, a comparative analysis (S60) is performed between the daily sleep curve (daily hypnogram) and the personalized standard sleep curve (Good Sleep hypnogram) for sleep data measured using the sleep care device (100) in the sleep evaluation device (300).

The above comparative analysis (S60) means comparing the degree of consistency between the sleep stage change of the personalized standard sleep curve (Good Sleep hypnogram) and the sleep stage change of the daily sleep curve (daily hypnogram).

At this time, the sleep length of the personalized standard sleep curve (Good Sleep hypnogram) and the sleep length and onset of sleep stages of the daily sleep curve (daily hypnogram) may be the same/similar or different from each other.

In the former case, that is, if the sleep length of the personalized standard sleep curve (Good Sleep hypnogram) and the sleep length and the start of the sleep stage of the daily sleep curve (daily hypnogram) are the same/similar, it is sufficient to use the DTW (Dynamic Time Warping) algorithm to calculate the similarity described below. The DTW (Dynamic Time Warping) algorithm is an algorithm that compares the similarity when the two values have different speeds (slopes) over time, and can be said to be a shape-based similarity calculation algorithm.

In the latter case, that is, if the sleep length or the starting point of the sleep stage of the personalized standard sleep curve (Good Sleep hypnogram) and the daily sleep curve (daily hypnogram) are different, a separate structure-based similarity calculation algorithm should be used to calculate the similarity between all states of the HMM, and the similarity calculation described below should be performed.

Subsequently, based on the above comparative analysis (S60), the sleep pattern similarity between the daily hypnogram and the personalized standard sleep curve is calculated, and a sleep index corresponding to the calculated sleep pattern similarity is determined (S70).

Sleep pattern similarity calculation is performed by 1) calculating the Fisher-Lao distance and Gini coefficient by considering the similarity probability between states included in the sleep pattern analysis probability model through the DTW (Dynamic Time Warping) algorithm or 2) the structure-based similarity calculation algorithm.

The above sleep index represents a sleep quality index corresponding to the calculation result of sleep pattern similarity. At this time, the size of the sleep quality index can be determined in proportion to the size of sleep pattern similarity.

Additionally, the sleep index is a numerical representation of the similarity between the daily hypnogram and the personalized standard sleep curve, and can be determined as a score value between 0 and 100.

The higher the similarity between the daily hypnogram and the personalized standard sleep curve, the higher the sleep index value. On the other hand, the lower the similarity between the daily hypnogram and the personalized standard sleep curve, the lower the sleep index value. For example, when the sleep index is 90, the similarity with the personalized standard sleep curve is higher than when the sleep index is 70, so the sleep quality can be evaluated as higher.

The sleep pattern similarity and sleep index produced in this manner can be output through a terminal (200) linked to a sleep care device (100) for visualization.

Alternatively, the sleep pattern similarity and sleep index may be output through the display of the sleep evaluation device (300).

Alternatively, if the operations according to the present embodiment implemented in the sleep evaluation device (300) are performed by executing an application installed in the terminal (200), the operations of FIG. 10 described above can be performed through communication between one or more servers/clouds/devices and the terminal (200), and the sleep evaluation results can be output through the terminal (200).

Meanwhile, in some embodiments, the conditions of the filtering (S40) can be expanded/reduced/changed through user input. Specifically, some embodiments of the present invention can expand the sleep data clustering model according to the selected filtering conditions, and compare and evaluate the user's sleep pattern similarity and sleep index with the group of the expanded clustering model. This will be described in more detail with reference to FIGS. 18A and 18B below.

In this way, the present invention enables more accurate sleep quality evaluation from the perspective of sleep structure changes by comparing the sleep stage changes while the user is sleeping with a personalized standard sleep curve generated by clustering sleep data with a high evaluation score, that is, when the user has had a good night's sleep. In addition, the comparison by sleep stage changes can be intuitively understood at a glance.

Below, FIGS. 11a and 11b are visualized example screens that allow the user to intuitively grasp at a glance the similarity of sleep patterns and the corresponding sleep index according to the change in sleep stages through a sleep evaluation device (300) or a terminal (200) according to an embodiment of the present invention. FIGS. 11a and 11b will be described as an example of what is shown through the display unit (225) of the terminal (200).

Referring to FIG. 11a, a first sleep wave corresponding to a personalized standard sleep curve is displayed in a first area (e.g., upper/left) (1100) of the display unit (225), and a second sleep wave corresponding to a daily sleep curve to be evaluated is simultaneously displayed in a second area (e.g., lower/right) (1120). In addition, a sleep index corresponding to the similarity (degree of agreement) between the first and second sleep waves (1100, 1120) may be displayed in a third area (e.g., lower display area) (1130) of the display unit (225).

The first sleep wave displayed in the first area (1100) is a sleep curve generated by learning by clustering days when the user slept well and by changing the sleep stage. The second sleep wave (1120) displayed in the second area (1120) is a sleep curve (target of evaluation) generated by the user sleeping for one day.

The sleep index displayed in the third area (1130) is a numerical value (%) indicating the degree of agreement between sleep stages when assuming that the second sleep wave overlaps the first sleep wave as shown in Fig. 11b. Since the sleep index is indicated as 50 in Fig. 11a, it can be said that the degree of agreement between the two sleep waves between sleep stages is 50%.

Depending on the embodiment, there may be multiple daily sleep curves generated during the day.

In this case, that is, when there are multiple daily sleep curves for the measured sleep data, the display unit (225) displays a second sleep wave corresponding to a daily sleep curve selected from among the multiple sleep curves in a manner that it overlaps the first sleep wave corresponding to the personalized standard sleep curve. For example, as illustrated in FIG. 11b, the first and second sleep waves (1110, 1121) are displayed in a first area (1100') of the display unit (225) in a manner that it overlaps the first and second sleep waves (1110, 1121), and Good Sleep Hypnoaram & Today's Hypnoaram indicating this is displayed. In addition, the second area (1120') of the display unit (225) displays sleep waves corresponding to the remaining daily sleep curves.

If no daily sleep curve among multiple sleep curves is selected, a daily sleep curve satisfying a set condition is designated as the selected daily sleep curve. Then, the sleep index is updated for the selected daily sleep curve. Here, the set condition may be one of the longest sleep time and the highest evaluation score.

In this case, when another daily sleep curve among multiple sleep curves is selected (e.g., through a touch gesture such as in FIG. 11b) through user input, etc., a second sleep wave corresponding to the selected daily sleep curve is displayed overlapping the first sleep wave, and the display of the sleep index is also changed.

Below, Fig. 12 is an example screen related to the data collection method for generating a personalized standard sleep curve.

Referring to FIG. 12, a user can request sleep data measurement or sleep evaluation (e.g., view sleep index) through the display unit (225) of a terminal (200) or a sleep evaluation device (300) linked to a sleep care device (100).

The illustrated screen (1200) may include measured total sleep time information (1201), a first input interface (1202) for continuing sleep care, a second input interface (1203) for an analysis request, and playback sound source information (1204) related to sleep induction/sleep care.

The first input interface (1202) may include a toggle function, in which case either 'continue sleep care' or 'stop sleep care' may be executed based on an input applied to the first input interface (1202). If 'continue sleep care' or 'stop sleep care' is repeated multiple times during the day, multiple daily sleep curves are generated for the number of repetitions.

Using the illustrated screen (1200), when sleep data measurement by the sleep care device (100) is completed, the measured sleep data is collected and the corresponding sleep evaluation is cumulatively stored (transmitted to the sleep evaluation device (300)).

The accumulation of sleep data and corresponding sleep evaluations is performed through the following process. First, a daily sleep curve (daily hypnogram) is generated based on sleep data measured through a sleep care device (100) (first process), and a sleep evaluation is performed on the generated daily sleep curve (second process). Next, sleep data and sleep evaluation corresponding to the generated daily sleep curve are stored (third process), and the first to third processes are repeated for a certain period of time (fourth process). At this time, the certain period of time may be, for example, 30 days, and the certain period of time may be variable.

Meanwhile, with respect to sleep data measured through the sleep care device (100) in the present invention, FIG. 14a shows a human sleep cycle, and FIG. 14b is an example showing a human brain wave measured by the sleep care device.

Referring to Figure 14a, when a person sleeps, he or she repeats a sleep cycle consisting of REM (rapid eye movement) sleep and NREM (non-rapid eye movement) sleep divided into four stages in a cycle of about 90 to 120 minutes, and usually has 3 to 5 sleep cycles per night. REM sleep refers to sleep in which rapid eye movement is observed, and NREM sleep refers to a stage in which the brain also falls asleep and eye movement does not occur, unlike REM sleep, and one falls into deep sleep. NREM sleep consists of shallow sleep stages 1 to 2 and deep sleep stages 3 to 4, and reaches the deepest stage of all sleep within the first 90 minutes of sleep, and stage 4, the deepest stage of sleep, can only be reached twice during 8 hours of sleep. The reason why the deep sleep stage is important is because it relieves physical fatigue of the entire body and regulates immune system function due to slow brain waves emitted in the deep sleep state. In particular, waking up from the deep sleep stage You will experience a state of confusion and confusion of intoxication.

As illustrated in FIG. 14a, the sleep stages of a user according to an embodiment of the present invention can be classified into a wake stage (WAKE), a shallow sleep stage (STAGE 1 to STAGE 2), a deep sleep stage (STAGE 3 to STAGE 4), and a REM sleep stage.

Human brain waves can be classified into delta waves (δ, 0.5–4 Hz), theta waves (θ, 4–8 Hz), alpha waves (α, 8–13 Hz), beta waves (β, 13–30 Hz), and gamma waves (γ, 30–50 Hz) depending on the frequency band. In addition, alpha waves (α, 8–13 Hz) and beta waves (β, 13–30 Hz) are mainly emitted during the waking stage, and while alpha waves decrease during the shallow sleep stage, theta wave (θ, 4–8 Hz) emission increases, and during the deep sleep stage, delta waves (δ, 0.5–4 Hz) can be emitted by more than 50%.

In the present invention, the daily hypnogram is an index/graph that represents sleep data corresponding to the user's biosignal measured through the sleep care device (100) according to changes in sleep stages. Therefore, the sleep curve of Fig. 14a may be an example of the daily hypnogram according to the present invention.

Continuing, with reference to FIG. 14b, the data processing unit of the sleep care device (100) or the data processing unit of the terminal (200) can receive the user's bio-signals, including the sensed user's brain wave signals and body movement signals, to determine the user's sleep state and classify the sleep stage.

Specifically, the data processing unit (of the sleep care device (100) or terminal (200)) can determine whether the user is awake (WAKE) or asleep (SLEEP) based on the body movement signal among the user's biosignals.

For example, if the body movement signal is greater than a preset threshold, the user can be determined to be awake. Also, if the body movement signal is less than the threshold, the user can be determined to be asleep. Next, if the user is determined to be asleep, the user's sleep stage can be specifically classified based on the brainwave signal. In other words, the user's sleep stage can be classified based on the ratio of the frequency bands that make up the brainwave signal.

Referring to Fig. 14b, it is possible to see how the distribution ratio of each frequency band constituting the user's brainwave signal changes over time. In other words, it is possible to analyze the change in the ratio of delta waves, theta waves, alpha waves, beta waves, and gamma waves included in the user's brainwave signal over time. Based on this, the current user's sleep stage can be classified into one of the light sleep stage, deep sleep stage, and REM sleep stage.

As mentioned above, the human brain generally mainly emits alpha waves (α, 8–13 Hz) and beta waves (β, 13–30 Hz) during the waking stage, alpha waves decrease during the shallow sleep stage while theta waves (θ, 4–8 Hz) increase in emission, and delta waves (δ, 0.5–4 Hz) are emitted by more than 50% during the deep sleep stage.

Therefore, by analyzing the change in the distribution ratio of each frequency band that constitutes the brainwave signal, it is possible to determine which sleep stage the user is currently in.

For example, if the delta wave area of the user's brainwave signal is more than 50%, the user's current sleep stage can be determined to be deep sleep, but it is not limited to this.

Meanwhile, a method of classifying a user's sleep stage through the user's body movement signal and brain wave signal can be automatically classified in real time through a machine learning algorithm. Specifically, a sleep stage classification model that classifies sleep stages can be trained by using brain wave signals and body movement signals as learning data by a learning processor among components of a sleep care device. That is, the user's brain wave signals and body movement signals can be trained as input data, and the frequency band-specific distribution ratio of the brain wave signals and body movement information can be trained as output data.

FIG. 13 is an exemplary flowchart illustrating a method for calculating a slip index according to an embodiment of the present invention.

Unless otherwise stated, it is to be noted that each process of the method illustrated in FIG. 13 can be performed by executing a specific application installed on a processor of a sleep evaluation device (300), a device/server/cloud/system, etc. that operates in conjunction with the sleep evaluation device (300), or a terminal (200) that is interlocked with the sleep care device (100).

Referring to FIG. 13, in the same manner as described in FIG. 12, sleep data measured through the sleep care device (100) is input and sleep evaluations for each are cumulatively stored (S1310). For example, measured sleep data and horizontal evaluations can be stored in the first database.

Subsequently, the sleep assessment generates a personalized standard sleep curve based on sleep data having results above a preset threshold (1320).

For example, if the score range of the sleep evaluation is from score 1 to score 5, sleep data satisfying the threshold score 5 (or score 4) is filtered as input data. At this time, the filtered input data, i.e., sleep data having a sleep evaluation higher than the threshold, can be stored in the second database.

According to an embodiment, the personalized standard sleep curve generated above can be stored in a separate storage device and continuously updated including sleep data having an evaluation score higher than a threshold value that is input thereafter. In this way, the personalized standard sleep curve is generated not only by using the clustered sleep data of the individual user, but also by only using sleep data with high quality of sleep according to sleep stage changes, so that customized comparative evaluation according to sleep stage changes is possible.

Meanwhile, according to an embodiment, generating a personalized standard sleep curve may include applying and training a sleep pattern analysis probability model to sleep data having results above a preset threshold.

In an embodiment of the present invention, a sleep pattern analysis probability model may use a Hidden Markov Model (HMM). At this time, a hidden state in the HMM may be defined as a plurality of sleep stages (e.g., wake, sleep 1/2/3/4, RAM).

When a personalized standard sleep curve (Good Sleep Hypnogram) is generated in this way, the sleep pattern similarity between the sleep data to be evaluated, i.e., the daily sleep curve (daily hypnogram), and the generated personalized standard sleep curve is calculated (S1330).

The calculation of sleep pattern similarity can be performed using a shape-based Dynamic Time Warping (DTW) algorithm or a separate structure-based algorithm.

The DTW (Dynamic Time Warping) algorithm can be used when the sleep length or the starting point of the change in the means stage of the daily sleep curve compared to the personalized standard sleep curve is the same or almost the same. The DTW (Dynamic Time Warping) algorithm has low complexity, but its accuracy is somewhat lower than that of the structure-based separate algorithm described below.

A separate algorithm based on structure can be used when the sleep length of the daily sleep curve is different and the starting point of each means stage change is different compared to the personalized standard sleep curve.

Specifically, the structure-based algorithm can be performed by calculating the Fisher-Lao distance considering the similarity probability between states (e.g., wake, sleep 1/2/3/4, RAM) included in the sleep pattern analysis probabilistic model (e.g., HMM), and calculating the Gini coefficient for calculating the correspondence considering the sparsity of the matrix.

In this way, a sleep index can be provided (output) based on the generated sleep pattern similarity (S1340).

Here, the sleep index is a numerical representation of the sleep quality index corresponding to the calculation result of the sleep pattern similarity. At this time, the sleep quality index is determined in proportion to the size of the sleep pattern similarity. For example, if the sleep pattern similarity is 80%, the sleep index can be expressed as '80'.

The above sleep index may be provided together with other sleep quality evaluation factors (e.g., awakening during sleep, time asleep, sleep efficiency, sleep time). In addition, the constant sleep index may be displayed on one screen simultaneously with each sleep wave corresponding to the daily sleep curve and the personalized standard sleep curve.

If a user sleeps multiple times a day, a daily sleep curve is generated for each sleep period, so multiple sleep curves may be displayed.

In such a case, a selected sleep curve among multiple sleep curves may be displayed with priority, and if not selected, a specific sleep curve satisfying preset conditions (e.g., longest sleep length, highest sleep evaluation score, most recently measured) may be displayed by default. The second sleep wave corresponding to the selected/default sleep curve may be displayed by overlapping the first sleep wave corresponding to the personalized standard sleep curve, thereby allowing for an intuitive understanding of the degree of sleep pattern consistency.

FIG. 15 illustrates a hidden Markov model as an example model related to the generation of a personalized standard sleep curve according to an embodiment of the present invention. And FIG. 16 is an example visualizing a standard sleep curve learned using the model of FIG. 15.

Referring to Figure 15, a hidden Markov model was used as a time series representation learning algorithm to model the characteristics of sleep patterns.

A Hidden Markov Model (HMM) is an extension of a Markov model that expresses the change of a phenomenon as a probability model, and is a probability model that focuses on inferring an unobservable hidden state through verifiable observation. In a Hidden Markov Model (HMM), state information is hidden and only the output information is observed.

In the present invention, the hidden state of the Hidden Markov Model (HMM) is defined as each stage of sleep (e.g., Wake, Sleep Stage N1/N2 (Light Sleep), Sleep Stage N3/N4 (Deep Sleep), REM), as shown in Fig. 15. In addition, HMM-based learning was performed by expressing this as the probability of a transition matrix.

At this time, the degree of model optimization can be identified through the log-likelihood function. In addition, by constructing a verification dataset for optimal model hyperparameter settings and proceeding with learning, the performance of the HMM model can be further improved.

In addition, in the present invention, in order to deal with the cyclical characteristics of sleep data (e.g., N1/N2/N3/N/REM cycles/repeats during sleep) while being robust to noise, a personalized standard sleep curve (Good Sleep Hypnogram) generated based on sleep data having a sleep evaluation score higher than a threshold among accumulated sleep data was learned through an HMM-based probabilistic model.

In addition, in the present invention, even after a personalized standard sleep curve (Good Sleep Hypnogram) is generated through HMM-based learning, HMM-based re-learning is continuously performed based on subsequently measured sleep data.

Figure 16 is an example of visualizing the result of learning personalized sleep stage changes through HMM, that is, a personalized standard sleep curve (Good Sleep Hypnogram).

In the present invention, a personalized standard sleep curve (Good Sleep Hypnogram) is an index/graph that represents a personalized standard sleep curve by sleep stage change based on clustering of sleep data with good sleep quality (i.e., having an evaluation score higher than a threshold) among sleep data corresponding to a user's bio-signal measured through a sleep care device (100). Therefore, FIG. 16 can be said to be an example of a personalized standard sleep curve (Good Sleep Hypnogram) according to an embodiment of the present invention.

This can be displayed simultaneously with the evaluated sleep data, i.e., the daily sleep curve (daily hypnogram) on the same screen. Alternatively, the daily sleep curve (daily hypnogram) can be displayed in a form in which it overlaps the visualized standard sleep curve (Good Sleep Hypnogram) so that the degree of consistency of the sleep pattern can be intuitively observed.

[Calculating sleep pattern similarity]

Meanwhile, the process of calculating the sleep pattern similarity between the daily sleep curve for measured sleep data and the personalized standard sleep curve is as follows.

Calculation of sleep pattern similarity between the above daily sleep curve and the above personalized standard sleep curve can be performed by one of two methods as follows.

First, when the length of sleep and/or the starting point of sleep stage changes are the same/almost similar, the sleep pattern similarity can be calculated using the Dynamic Time Warping (DTW) algorithm.

Second, when the length of sleep is different and the starting point of sleep stage change is also different, the DTW (Dynamic Time Warping) algorithm is not suitable for comparing time series data and the accuracy of similarity calculation is low. Therefore, in such cases, the sleep pattern similarity can be calculated using a structure-based similarity calculation algorithm.

The structure-based similarity calculation algorithm is performed by calculating the Fisher-Rao distance and Gini coefficient considering the similarity probability between states included in the sleep pattern analysis probability model (e.g., HMM) according to the present invention.

Specifically, the structure-based similarity calculation algorithm can be expressed as a matrix as shown in mathematical formula 1 through the Fisher-Rao distance, which considers the probability of the similarity between all states of each model.

Here, S _ε represents the Fisher-Rao probability distance, λ and π represent HMM and stationary distribution, respectively, and v _i represents the ith state.

Afterwards, in order to calculate the degree of agreement considering the sparsity of the matrix, the Gini coefficient is calculated as in the following mathematical expression 2 to derive the similarity.

Here, r and c represent rows and columns, respectively, and H() represents the Gini coefficient, which is an indicator of impurity.

Meanwhile, since the DTW (Dynamic Time Warping) algorithm is capable of fast time score estimation, it will be possible to calculate more accurate similarity by using it in parallel with the aforementioned structure-based similarity calculation algorithm or by cross-validating it with the structure-based similarity calculation algorithm.

Next, FIG. 17a is an example screen for comparing the records of a user's sleep index according to an embodiment of the present invention, and FIG. 17b is an example screen displayed on a display unit of a terminal (200) or a sleep evaluation device (300) when requesting the recording of a sleep index before generating a personalized standard sleep curve.

In step S1340 of Fig. 13, when providing a sleep index, a user interface for displaying the sleep index on a daily, weekly, or monthly basis may be provided together. To this end, a sleep index corresponding to a sleep pattern similarity between a personalized standard sleep curve (Good Sleep Hypnogram) and a daily sleep curve (daily hypnogram) may be recorded and stored on a daily, weekly, or monthly basis.

The user interface (1411) may be an input means for selectively comparing daily (1710), weekly (1720), and monthly (1730) records for the sleep index, as illustrated in FIG. 17a.

When the user interface (1411) is positioned at daily (1710), sleep indices for sleep data for each day are provided, and when it is positioned at weekly (1720), average records (e.g., numerical changes and average values) for sleep indices for each week are provided together. In addition, when the user interface (1411) is positioned at monthly (1730), average records (e.g., numerical changes and average values) for sleep indices for each month are provided on one screen.

Referring to Fig. 17a, the sleep index for days 4.4 to 4.25 is displayed by day, and for example, when day 4.8 is selected, the sleep index (91.5%) for that day and the displayed daily sleep index average (91.3%) can be displayed on the sleep index record screen (1750). That is, the daily records of the sleep index can be compared.

Meanwhile, if sleep data has not been accumulated enough to check the sleep index record, an indication that sleep data is being acquired may be provided.

Specifically, in response to receiving a sleep analysis request prior to the step of generating a personalized standard sleep curve, the sleep assessment device (300) may determine that a personalized standard sleep curve is not generated, and in response to the determination, provide visual information via the display unit indicating that sleep data is being collected to provide the sleep index.

In such cases, only the daily sleep curve corresponding to the measured sleep data may be visually displayed, or the similarity with the clustered standard sleep curve of the group entered by the user as the initial setting may be displayed. In the latter case, information (e.g., a pop-up message) may be displayed together indicating that the comparison is with the clustered standard sleep curve provided by default, not the personalized standard sleep curve.

For example, as illustrated in FIG. 17b, visual information (1740) such as “collecting sleep index” may be displayed, and only the average record (1760) of the currently accumulated sleep index may be provided.

Next, FIGS. 18a and 18b are example screens displayed on the display unit of a terminal (200) or a sleep evaluation device (300) when a sleep analysis request is made. Referring to these, a method for visually providing the results of comparing the sleep index of a group corresponding to a selected filtering condition and the sleep index of a user will be described.

In response to a user's sleep analysis request, the user's own sleep index can be compared with that of others/groups.

To this end, according to an embodiment of the present invention, when providing a sleep index corresponding to the result of calculating sleep pattern similarity between a personalized standard sleep curve (Good Sleep Hypnogram) and a daily sleep curve (daily hypnogram), one or more related sleep evaluation factors can be provided by comparing them with the sleep index of the first group corresponding to the selected filter item.

Here, the sleep evaluation factors may include, for example, age, gender, occupation, BMI index, etc. as filtering items for the control group to compare the sleep index. At this time, the setting value of the filtering item may be set as an initial value selected through user input. For example, if the user is a man in his 30s and his occupation is a freelancer, the corresponding value is set as an initial setting value through user input.

Filtering items can be none, one, or multiple selected.

If no filtering item is selected, for example, as illustrated in FIG. 18a, the record compared to one's recent sleep index can be displayed as visual information (e.g., in graph form) (1801). Alternatively, although not illustrated, visual information can be provided for the record compared to one's specific day/specific period average sleep index.

When more than one filtering item is selected, as shown in Fig. 18b, this can be performed by inputting a selection for one or more of the presented filtering items (1810, 1820, 1830, 1840), for example, 30s, male, occupation (e.g., freelancer), BMI 30.

In this case, the sleep index is updated based on the changed filter items according to the selection input. For example, since the control group changes depending on the changed filter items, the value of the sleep index corresponding to the sleep pattern similarity also changes.

The selected filter items, such as age (e.g., 30s), gender (e.g., male), occupation (e.g., freelancer), BMI (e.g., 30), etc., are transmitted to the sleep evaluation device (300), and the sleep evaluation device (300) recalculates the user's sleep index based on the data filtered by the transmitted conditions.

To this end, the sleep evaluation device (300) provides filtering data corresponding to the input keyword so that artificial intelligence (AI) can automatically illustrate based on the recognized keyword (e.g., 30s, male, freelancer) according to the above selection input.

The resulting values can be transmitted to the terminal (200) (e.g., via an application) and displayed on the display unit or output through the display unit of the sleep evaluation device (300).

According to an embodiment, the updated sleep index and one or more sleep evaluation factors related thereto are provided in the form of a comparison with a second group of sleep indices corresponding to filter items changed according to the selection input. For example, as illustrated in FIG. 18b, evaluation factor items related to the comparison graph (e.g., falling asleep time, sleeping time, sleep efficiency, wakefulness during sleep, falling asleep + sleeping time, etc.) may be expressed in numerical values.

In some embodiments, the initial settings of the filtering items can be changed. In this case, the calculation complexity for calculating the sleep pattern similarity and updating the corresponding sleep index will increase significantly, but the user can change the filtering items to the desired settings to compare his or her sleep quality. For example, the user is a man in his 30s and is a freelancer, but the filtering item for the control group can be selected as a 'woman' in her 30s for comparison.

As described above, according to an embodiment of the present invention, by clustering sleep data evaluated as having good sleep quality among the user's sleep data and utilizing a customized personalized standard sleep curve (Good Sleep Hypnogram) as a comparison standard based on the clustered sleep data, the user's sleep evaluation is performed according to the clustered customized sleep stage change, thereby enabling a more accurate evaluation of sleep quality. In addition, the similarity according to the sleep stage change for the daily sleep curve (Daily Hypnogram) to be compared with the customized standard sleep curve (Good Sleep Hypnogram) can be intuitively compared and identified at a glance. In addition, the user's own sleep evaluation can be easily checked for a desired period, and the results compared to other groups under selected conditions can be checked simultaneously.

The above-described present disclosure can be implemented as a computer-readable code on a medium having a program recorded thereon. The computer-readable medium includes all kinds of recording devices that store data that can be read by a computer system. Examples of the computer-readable medium include a hard disk drive (HDD), a solid state disk (SSD), a silicon disk drive (SDD), a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like. Furthermore, the computer may include a control unit/processor of a sleep evaluation device.

Further scope of applicability of the present invention will become apparent from the detailed description below. However, since various changes and modifications within the spirit and scope of the present invention will become apparent to those skilled in the art, it should be understood that the detailed description and specific embodiments, such as preferred embodiments of the present invention, are given by way of example only.

The features, structures, effects, etc. described in the embodiments above are included in at least one embodiment of the present invention, and are not necessarily limited to just one embodiment. Furthermore, the features, structures, effects, etc. exemplified in each embodiment can be combined or modified and implemented in other embodiments by a person having ordinary knowledge in the field to which the embodiments belong. Therefore, the contents related to such combinations and modifications should be interpreted as being included in the scope of the present invention.

In addition, although the above description has been made with reference to examples, these are merely examples and do not limit the present invention, and those with ordinary knowledge in the field to which the present invention pertains will recognize that various modifications and applications not exemplified above are possible without departing from the essential characteristics of the present invention. For example, each component specifically shown in the examples can be modified and implemented. And the differences related to such modifications and applications should be interpreted as being included in the scope of the present invention defined in the appended claims.

Claims (20)

A step of inputting sleep data measured through a sleep care device and accumulatively storing sleep evaluations; A step of generating a personalized standard sleep curve based on sleep data having a sleep evaluation above a preset threshold; A step of calculating sleep pattern similarity between the daily sleep curve for measured sleep data and the personalized standard sleep curve; and Including a step of visually providing the generated sleep pattern similarity and the corresponding sleep index, A method for assessing sleep using sleep stage change patterns. In the first paragraph, The above slip index is, The sleep quality index corresponding to the result of calculating the sleep pattern similarity is determined as a size value proportional to the sleep pattern similarity. A method for assessing sleep using sleep stage change patterns. In the first paragraph, The steps provided visually above are: Including a step of simultaneously displaying each sleep wave and the sleep index corresponding to the daily sleep curve and the personalized standard sleep curve for the measured sleep data on one screen. A method for assessing sleep using sleep stage change patterns. In the third paragraph, If there are multiple daily sleep curves for the above measured sleep data, The steps provided visually above are: A step of displaying all of the above multiple daily sleep curves; A step of displaying a sleep wave corresponding to a daily sleep curve selected from among the plurality of daily sleep curves by overlapping it with a sleep wave corresponding to the personalized standard sleep curve, A method for assessing sleep using sleep stage change patterns. In paragraph 4, The steps provided visually above are: Including a step of displaying a sleep index for a daily sleep curve satisfying a preset condition among the above multiple daily sleep curves. A method for assessing sleep using sleep stage change patterns. In paragraph 5, Including a step of changing the sleep index for the daily sleep curve satisfying the above preset condition to the sleep index for the selected daily sleep curve and displaying it. A method for assessing sleep using sleep stage change patterns. In the first paragraph, The step of cumulatively storing the above sleep evaluation is: The first step is to create a daily sleep curve based on sleep data measured through a sleep care device; The second step is to perform a sleep evaluation on the generated daily sleep curve; A third process of storing sleep data and sleep evaluation corresponding to the generated daily sleep curve; and Including a fourth process of repeating the first to third processes for a certain period of time. A method for assessing sleep using sleep stage change patterns. In the first paragraph, The steps for generating the above personalized standard sleep curve are: A step of generating sleep data by applying a sleep pattern analysis probability model to sleep data having a sleep evaluation exceeding the preset threshold. A method for assessing sleep using sleep stage change patterns. In Article 8, The steps for generating the above personalized standard sleep curve are: The sleep pattern analysis probability model includes a step of performing learning on sleep data having a sleep evaluation greater than the threshold by applying a Hidden Markov Model (HMM) that defines multiple sleep stages as states. A method for assessing sleep using sleep stage change patterns. In Article 8, The step of calculating the above sleep pattern similarity is: This is the step that is calculated using the DTW (Dynamic Time Warping) algorithm. A method for assessing sleep using sleep stage change patterns. In Article 8, The step of calculating the above sleep pattern similarity is: This is a step of calculating using a structure-based similarity calculation algorithm that is performed by calculating the Fisher-Lao distance and the Gini coefficient considering the similarity probability between states included in the above sleep pattern analysis probability model. A method for assessing sleep using sleep stage change patterns. In the first paragraph, The steps provided visually above are: A step of providing a user interface for displaying the calculated sleep pattern similarity and the corresponding sleep index on a daily, weekly, or monthly basis, A method for assessing sleep using sleep stage change patterns. In the first paragraph, A step of receiving a sleep analysis request prior to the step of generating the personalized standard sleep curve; In response to the sleep analysis request, determining that the personalized standard sleep curve is not generated; and In response to the above decision, a step of displaying visual information indicating that sleep data is being collected to provide the sleep index is included. A method for assessing sleep using sleep stage change patterns. In the first paragraph, The steps provided visually above are: Further comprising a step of providing a sleep index and one or more sleep evaluation factors related thereto by comparing them with a sleep index of a first group corresponding to a selected filter item. A method for assessing sleep using sleep stage change patterns. In Article 14, In the step of providing the above comparison, A step of receiving an input for changing the above filter item; A step of updating the sleep index based on the changed filter items according to the above input; and Comprising a step of providing the updated sleep index and one or more sleep evaluation factors related thereto by comparing them with the sleep index of the second group corresponding to the filter items changed according to the input. A method for assessing sleep using sleep stage change patterns. A communication unit that receives the user's sleep data measured through a sleep care device; A data processing unit that accumulates and stores sleep evaluations for the measured sleep data in a storage device and generates a personalized standard sleep curve based on sleep data having a sleep evaluation exceeding a preset threshold; and Including a display section, The above data processing unit, Calculate the sleep pattern similarity between the daily sleep curve for the above measured sleep data and the generated personalized standard sleep curve, Controlling the display unit to visually display the calculated sleep pattern similarity and the corresponding sleep index, Sleep assessment device. In Article 16, The above slip index is, The sleep quality index corresponding to the result of calculating the sleep pattern similarity is determined as a size value proportional to the sleep pattern similarity. Sleep assessment device. In Article 16, The above display part, Simultaneously displaying the sleep index and each sleep wave corresponding to the daily sleep curve and the personalized standard sleep curve for the measured sleep data on one screen. Sleep assessment device. In Article 16, The above data processing unit, Applying a sleep pattern analysis probability model to sleep data having a sleep evaluation exceeding the preset threshold value to generate the personalized standard sleep curve. Sleep assessment device. In Article 19, The above data processing unit, As the sleep pattern analysis probability model, a Hidden Markov Model (HMM) that defines multiple sleep stages as states is applied to learn about sleep data having a sleep evaluation higher than the threshold. Sleep assessment device.

Images