JP2010148575A - Apparatus and method for determining sleep stage - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a sleep stage determination apparatus and a sleep stage determination method for determining an accurate sleep stage. <P>SOLUTION: A boundary value change part 25 calculates the boundary value of a sleep stage within a range where different sleep stage distributions overlap with each other for a biological feature amount in the sleep stage distributions which are the distributions of the biological feature amounts for each sleep stage of an individual for the biological feature amount regarding the sleep of the individual, and a sleep stage determination part 34 determines the sleep stage of the individual on the basis of the biological feature amount and the boundary value. Thus, compared to the method of determining the sleep stage by a predetermined threshold, the more accurate sleep stage is determined. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a sleep stage determination apparatus and a sleep stage determination method, and more particularly, to a sleep stage determination apparatus and a sleep stage determination method for acquiring a biometric feature related to an individual's sleep and determining a solid sleep stage.

2. Description of the Related Art Conventionally, there has been proposed an apparatus that determines a sleep stage of an individual by detecting a biometric feature related to the sleep of the individual. For example, Patent Document 1 includes a biological information sensor that detects biological information, and a heart rate, a standard deviation of the heart rate, a respiratory rate, a standard deviation of the respiratory rate, and a body motion number from an output signal of the biological information sensor. The sleep depth estimation circuit comprising: a biological information processing circuit for detecting a plurality of sleep depth basic data; and a sleep depth estimation circuit for repeating the sleep depth estimation processing based on the plurality of sleep depth basic data obtained from the biological information processing circuit. Discloses a sleep depth estimation apparatus that executes an estimation process combining discriminant analysis and a state transition probability model on a plurality of sleep depth basic data obtained from a biological information processing circuit.
Japanese Patent No. 3658580

  However, since the heart rate, respiration, and the like fluctuate depending on the season, physical condition, and the like in the above technique, there is a problem that the sleep stage determined with a preset threshold value is not necessarily an accurate sleep stage of the individual. In this case, it is conceivable to change the threshold for determining the sleep stage before using the device according to various conditions such as the age of the individual. However, even in a specific individual, the quality of sleep can vary greatly depending on the season such as summer and autumn. Furthermore, even in a specific individual, depending on the physical condition, there is a possibility that the quality of sleep may change slightly on a weekly basis. Therefore, there is a demand for a method that can accurately estimate the sleep stage based on the individual characteristics automatically using the biometric features related to the sleep of the individual.

  This invention is made | formed in view of such a situation, The objective is to provide the sleep stage determination apparatus and sleep stage determination method which can perform the determination of a more accurate sleep stage.

  The present invention determines the sleep stage of an individual from the degree of overlapping of different sleep stage distributions with respect to the biometric feature value in the sleep stage distribution that is the distribution of the biometric feature value for each sleep stage with respect to the biometric feature value related to the sleep of the individual. It is a sleep stage determination apparatus provided with the sleep stage determination means.

  According to this configuration, in the sleep stage distribution that is the distribution of the biometric feature amount for each individual sleep stage with respect to the biometric feature amount related to the individual sleep, the sleep stage determination unit includes different sleep stage distributions for the biometric feature amount. Since an individual's sleep stage is determined from the degree of overlap, it is possible to perform a more accurate determination of the sleep stage as compared to a method of determining a sleep stage using a predetermined threshold.

  Further, the present invention relates to a sleep stage distribution, which is a distribution of biometric features for each sleep stage relative to a biometric feature related to an individual's sleep, and determines an individual's sleep stage from the degree to which different sleep stage distributions overlap with each other. It is a sleep stage determination apparatus provided with the sleep stage determination means to determine, and the travel assistance means which provides the information regarding the driving | running return of an individual to the vehicle based on the sleep stage of the individual determined by the sleep stage determination means.

  According to this configuration, in the sleep stage distribution, which is the distribution of the biometric feature amount for each sleep stage with respect to the biometric feature amount related to the individual sleep, the sleep stage of the individual is determined based on the degree of overlap of the different sleep stage distributions with respect to the biometric feature amount. Since the determination and the provision of information regarding the return of the vehicle to the vehicle are performed, it is possible to perform safe driving after sufficient rest.

  In the sleep stage distribution, which is a distribution of biometric features for each sleep stage with respect to biometric features relating to an individual's sleep, the present invention is within a range in which different sleep stage distributions overlap each other with respect to the biometric features. A sleep stage determination device comprising boundary value calculation means for calculating a boundary value, and sleep stage determination means for determining the sleep stage of an individual based on a biometric feature and the boundary value.

  According to this configuration, in the sleep stage distribution that is the distribution of the biometric feature amount for each individual sleep stage with respect to the biometric feature amount related to the individual sleep, the boundary value calculation means includes different sleep stage distributions for the biometric feature amount. A method for determining a sleep stage based on a predetermined threshold in order to calculate a sleep stage boundary value within the overlapping range and the sleep stage determination means to determine the individual sleep stage based on the biometric feature value and the boundary value. Compared to the above, it becomes possible to determine the sleep stage more accurately.

  In this case, the biometric feature amount acquisition means for acquiring the biometric feature amount related to the sleep of the individual and the sleep that is the distribution of the biometric feature amount for each sleep stage of the unspecified number of individuals with respect to the biometric feature amount related to the sleep of the unspecified number of individuals The stage distribution further includes an initial boundary value setting unit that sets an initial boundary value of the sleep stage within a range in which different sleep stage distributions overlap with the biometric feature amount, and the sleep stage determination unit includes the biometric feature amount It is preferable to determine the sleep stage of the individual based on the biometric feature amount acquired by the acquisition unit and at least one of the initial boundary value set by the initial boundary value setting unit and the boundary value calculated by the boundary value calculation unit. is there.

  According to this configuration, the biometric feature amount acquisition unit acquires the biometric feature amount related to the sleep of the individual, and the initial boundary value setting unit performs sleep of the unspecified number of individuals with respect to the biometric feature amount related to the sleep of the unspecified number of individuals. In the sleep stage distribution, which is a distribution of biometric features for each stage, an initial boundary value of the sleep stage is set within a range where different sleep stage distributions overlap with the biometric features, and the sleep stage determination means Since the individual's sleep stage is determined based on the biometric feature amount acquired by the amount acquisition means and at least one of the initial boundary value set by the initial boundary value setting means and the boundary value calculated by the boundary value calculation means. It becomes possible to determine the sleep stage more accurately based on a large amount of data.

  In this case, the boundary value calculation means is a biometric feature in the sleep stage distribution, which is a distribution of biometric features for each sleep stage of the unspecified number of individuals with respect to the biometric feature value related to sleep of the unspecified number of individuals. In a range in which different sleep stage distributions overlap with each other, the initial value according to the number of sample values of the biometric feature amount acquired by the biometric feature amount acquisition unit included in each range divided into two by the initial boundary value The sleep stage determination unit calculates the boundary value by correcting the boundary value, and the sleep stage determination unit converts the biometric feature amount acquired by the biometric feature amount acquisition unit and the boundary value calculated by the boundary value calculation unit by correcting the initial boundary value. It is preferred to determine the sleep stage of the individual based on it.

  According to this configuration, the boundary value calculation means is configured to calculate the biometric feature amount in the sleep stage distribution that is the distribution of the biometric feature amount for each sleep stage of the unspecified large number of individuals with respect to the biometric feature amount related to the sleep of the unspecified large number of individuals. In the range where the different sleep stage distributions overlap, the initial boundary value is determined according to the number of sample values of the biometric feature amount acquired by the biometric feature amount acquisition means included in each range divided into two by the initial boundary value. The boundary value is calculated by correction, and the sleep stage determination unit is configured to detect the individual based on the biometric feature amount acquired by the biometric feature amount acquisition unit and the boundary value calculated by the boundary value calculation unit by correcting the initial boundary value. Since the sleep stage is determined, each time the individual sleeps, the data for determining the sleep degree reflecting the characteristics of the individual is updated, and the influence of heart rate variability due to the season, physical condition change and the like can be eliminated.

  In this case, the boundary value calculating means sleeps differently with respect to the biological feature amount in the sleep stage distribution, which is a distribution of the biological feature amount for each sleep stage of the unspecified number of individuals with respect to the biological feature amount related to the sleep of the unspecified number of individuals. In the range where the step distributions overlap, the center of gravity of each range divided into two by the initial boundary value is calculated according to the sample value of the biometric feature amount included in each range divided into two by the initial boundary value. It is preferable that the boundary value is calculated by correcting the initial boundary value according to the number of sample values of the biometric feature amount included between each centroid and the initial boundary value.

  According to this configuration, the boundary value calculation means is configured to calculate the biometric feature amount in the sleep stage distribution that is the distribution of the biometric feature amount for each sleep stage of the unspecified large number of individuals with respect to the biometric feature amount related to the sleep of the unspecified large number of individuals. In the range where different sleep stage distributions overlap, the center of gravity of each range divided into two by the initial boundary value according to the sample value of the biometric feature amount included in each range divided into two by the initial boundary value And the boundary value is calculated by correcting the initial boundary value according to the number of sample values of the biometric feature amount included between each centroid and the initial boundary value, so that finer correction is possible. .

  On the other hand, the sleep stage distribution further includes a reliability calculation unit that calculates the reliability of the sleep stage with respect to the biometric feature amount according to the degree to which the different sleep stage distributions overlap with the biometric feature amount. It is preferable to determine the sleep stage of the individual based on the reliability calculated by the reliability calculation means.

  According to this configuration, the reliability calculation means calculates the reliability of the sleep stage with respect to the biometric feature according to the degree of overlap between the sleep stage distributions different from the biometric feature in the sleep stage distribution, and the sleep stage Since the determination means determines the sleep stage of the individual based on the reliability calculated by the reliability calculation means, the accuracy of the sleep stage determination is increased by using data relating to the reliability of the sleep stage with respect to the biometric feature amount. Is possible.

  In this case, it is preferable that the reliability calculation unit calculates the reliability lower as the degree of overlapping of the sleep stage distributions increases, and calculates the reliability higher as the degree of the overlap of sleep stage distributions decreases.

  In the sleep stage distribution, the greater the degree of overlap between the sleep stage distributions, the harder it is to determine which sleep stage is, so the reliability is lower, and the lower the degree of overlap between the sleep stage distributions, Since it is easy to determine whether or not there is a higher reliability, this configuration makes it possible to accurately determine the reliability of the sleep stage with respect to the biological feature.

  In this case, the biometric feature acquisition unit acquires each of a plurality of types of biometric feature amounts of the individual, and the reliability calculation unit determines a sleep stage different from the biometric feature amount in each sleep stage distribution for each type of biometric feature amount. The sleep stage reliability for each type of biometric feature is calculated according to the degree to which the distributions overlap, and the sleep stage determination means is reliable for each type of biometric feature acquired by the biometric feature acquisition means. It is preferable to determine the sleep stage of the individual based on the sleep stage having the highest reliability determined by the degree determination means.

  According to this configuration, the biometric feature acquisition unit acquires each of a plurality of types of biometric feature amounts of the individual, and the reliability calculation unit determines the biometric feature amount for each sleep stage distribution for each type of biometric feature amount. Depending on the degree of overlap between the different sleep stage distributions, the reliability of the sleep stage for each type of biometric feature is calculated, and the sleep stage determination means uses each type of biometric feature acquired by the biometric feature acquisition means. On the other hand, in order to determine the sleep stage of an individual based on the sleep stage with the highest reliability determined by the reliability determination means, the sleep stage determination based on each type of biometric feature has the highest reliability. As a result, it becomes possible to more accurately determine the sleep stage that further matches the characteristics of the individual.

  On the other hand, the biometric feature is at least one of an individual's heartbeat and heartbeat fluctuation, and the sleep stage is preferably acquired based on the individual's brain wave.

  According to this configuration, since it is possible to obtain heart rate data such as heartbeats and heartbeat fluctuations of an individual capable of acquiring biometric features in a short time, and to acquire the sleep stage accurately by acquiring the individual's brain waves, more immediate This makes it possible to determine the sleep level with excellent performance and accuracy.

  On the other hand, according to the present invention, in the sleep stage distribution that is the distribution of the biometric feature amount for each individual sleep stage with respect to the biometric feature amount related to the sleep of the individual, the sleep of the individual is determined based on the degree of overlap between the different sleep stage distributions for the biometric feature amount. It is a sleep stage determination method including the process of determining a stage.

  The present invention also relates to the sleep stage distribution, which is the distribution of the biometric feature amount for each individual sleep stage relative to the biometric feature amount related to the sleep of the individual, from the degree to which the different sleep stage distributions overlap with the biometric feature amount. It is a sleep stage determination method including a step of determining a stage and a step of providing information related to the return of driving of the individual to the vehicle based on the determined sleep stage of the individual.

  Furthermore, the present invention relates to a sleep stage distribution, which is a distribution of biometric features for each sleep stage of an individual with respect to a biometric feature related to the sleep of the individual, within a range in which different sleep stage distributions overlap with respect to the biometric feature. A sleep stage determination method including a step of calculating a boundary value of a stage and a step of determining a sleep stage of an individual based on a biometric feature value and the boundary value.

  In this case, in the step of acquiring the biometric feature amount related to the sleep of the individual and the sleep stage distribution which is the distribution of the biometric feature amount for each sleep stage of the unspecified large number of individuals with respect to the biometric feature amount related to the sleep of the unspecified large number of individuals, A step of setting an initial boundary value of the sleep stage within a range in which different sleep stage distributions overlap with the biometric feature amount, and the acquired biometric feature amount and at least one of the initial boundary value and the boundary value It is preferred to determine the sleep stage of the individual based on it.

  In this case, the range in which different sleep stage distributions overlap with the biometric feature amount in the sleep stage distribution, which is the distribution of the biometric feature amount for each sleep stage of the unspecified large number of individuals with respect to the biometric feature amount related to sleep of the unspecified large number of individuals And calculating the boundary value by correcting the initial boundary value according to the number of sample values of the acquired biometric feature amount included in each range divided into two by the initial boundary value, and the acquired biometric feature amount It is preferable to determine the sleep stage of the individual based on the boundary value calculated by correcting the initial boundary value.

  In this case, the range in which different sleep stage distributions overlap with the biometric feature amount in the sleep stage distribution, which is the distribution of the biometric feature amount for each sleep stage of the unspecified large number of individuals with respect to the biometric feature amount related to sleep of the unspecified large number of individuals The center of gravity of each range divided into two by the initial boundary value is calculated according to the sample value of the biometric feature amount included in each range divided into two by the initial boundary value. It is preferable to calculate the boundary value by correcting the initial boundary value according to the number of sample values of the biometric feature amount included between the values.

  On the other hand, the sleep stage distribution further includes a step of calculating the reliability of the sleep stage with respect to the biometric feature amount according to the degree of overlap between the sleep stage distributions different from each other with respect to the biometric feature amount. It is preferred to determine the stage.

  In this case, it is preferable to calculate the reliability lower as the degree of overlapping of the sleep stage distributions increases, and to calculate the reliability higher as the degree of the overlapping of sleep stage distributions decreases.

  In this case, each of a plurality of types of biometric features of the individual is acquired, and in each sleep stage distribution for each type of biometric feature, according to the degree to which different sleep stage distributions overlap with the biometric feature, The reliability of the sleep stage with respect to the biometric feature is calculated, and the sleep stage of the individual is determined based on the sleep stage having the highest reliability with respect to each type of biometric feature acquired by the biometric feature acquisition unit. Is preferred.

  On the other hand, the biometric feature is at least one of an individual's heartbeat and heartbeat fluctuation, and the sleep stage is preferably acquired based on the individual's brain wave.

  According to the sleep stage determination device and the sleep stage determination method of the present invention, it is possible to perform a more accurate determination of the sleep stage.

  Embodiments of the present invention will be described below with reference to the accompanying drawings.

  FIG. 1 is a block diagram illustrating a configuration of a sleep stage determination device 1 according to the present embodiment. The sleep stage determination apparatus 1 of the present embodiment is configured as an in-vehicle apparatus as an example, but is not limited to in-vehicle use, and is for determining a sleep stage (an international standard by Rechtschaffen and Kales) of a subject user. belongs to.

  As shown in FIG. 1, the sleep stage determination device 1 of this embodiment includes a module A, a module B, and a module C. The module C is connected to a driving support ECU (driving support means) 100 that provides driving support to the user who is a driver based on the result of the sleep stage determination by the module C. Module A may be constructed as a database outside the apparatus, and module B and module C are built in the apparatus.

  As shown in FIG. 2, the module A calculates the initial boundary value and reliability of the sleep stage of each index, which are various biometric features in an unspecified number of nap data. In module B, the nap data of the user as the subject is acquired, and the boundary value of the sleep stage of each index is changed from the acquired data. In module C, the sleep level of the user who is the subject is determined based on the reliability with respect to the boundary value to which the reliability of each index is given.

  Returning to FIG. 1, the module A includes a general nap data acquisition unit 11, a histogram creation unit 12, an initial boundary setting unit 13, a changed section setting unit 14, and a reliability calculation unit 15. The general nap data acquisition unit 11 is for acquiring nap data of an unspecified number of generally healthy people. The nap data in this case indicates sleep stages of an unspecified number of people with respect to index values such as brain waves, pulse waves (heart rate fluctuation), heart rate, body temperature, and respiration.

  The histogram creation unit 12 is for creating a sleep stage histogram of each index from the nap data acquired by the general nap data acquisition unit 11. The initial boundary setting unit (initial boundary value setting means) 13 is for setting the initial boundary value of the sleep stage on the histogram of each index created by the histogram creation unit 12. The change interval setting unit 14 is for setting a change interval of the sleep stage on the histogram of each index created by the histogram creation unit 12. The reliability calculation unit (reliability calculation means) 15 is for calculating the reliability of the sleep stage determined by the histogram of each index created by the histogram creation unit 12.

  The module B includes a user nap data acquisition unit 21, a boundary value / change section acquisition unit 22, a centroid calculation unit 23, a mass ratio calculation unit 24, and a boundary value change unit 25. The user nap data acquisition unit (biometric feature amount acquisition means) 21 is for acquiring nap data of a user who is a subject. Specifically, the user nap data acquisition unit 21 can acquire brain waves, pulse waves (heart rate fluctuation), heart rate, body temperature, respiration, and the like as indices of nap data. The boundary value / change interval acquisition unit 22 is for acquiring the boundary value and change interval of each index from the supplementary boundary value setting unit 13 and the change interval setting unit 14 of the module A.

  The center-of-gravity calculation unit (boundary value calculation means) 23 calculates the center of gravity of the nap data of each index acquired by the user nap data acquisition unit 21 in each range divided into two by the boundary value of the change section of the histogram of each index. It is for calculating. The mass ratio calculation unit (boundary value calculation means) 24 calculates a ratio between the mass of the nap data on one side of the boundary value and the mass of the nap data on the other side in the changed section. It is. The boundary value changing unit (boundary value calculating means) 25 is for changing the boundary value based on the mass ratio of the nap data calculated by the mass ratio calculating unit 24.

  The module C includes a user nap data acquisition unit 31, a boundary value / change interval acquisition unit 32, a reliability comparison unit 33, and a sleep stage determination unit 34. The user nap data acquisition unit (biometric feature amount acquisition means) 31 is for acquiring nap data of a user who is a subject. Specifically, the user nap data acquisition unit 31 can acquire an electroencephalogram, a pulse wave (heart rate fluctuation), a heart rate, a body temperature, a breath, and the like as indexes of the nap data. Since the boundary value is obtained, for example, it is possible to acquire less types of data than the user nap data acquisition unit 21 of the module B, such as omitting the detection of brain waves. Alternatively, the user nap data acquisition unit 21 of the module B and the user nap data acquisition unit 31 of the module C may be shared.

  The boundary value / change interval acquisition unit 32 is for acquiring the boundary value and change interval of each index from the initial boundary setting unit 13 and the change interval setting unit 14 of the module A and the boundary value change unit 25 of the module B. is there. The reliability comparison unit 33 is for comparing the reliability of the sleep stage according to each index calculated by the reliability calculation unit 15 of the module A. The sleep stage determination unit (sleep stage determination means) 34 finally determines the user's sleep stage based on the sleep stage of the user who is the subject determined by the boundary value of each index and the reliability of each index. Is to do.

  The driving support ECU 100 is for controlling the vehicle by driving an accelerator actuator, a brake actuator, and a steering actuator (not shown) based on the sleep stage determination result of the user who is the driver by the module C. Further, the driving support ECU 100 provides and notifies information for supporting the user through a display, a speaker, and the like (not shown) based on the result of the sleep stage determination of the user who is the driver by the module C.

  Hereinafter, operation | movement of the sleep stage determination apparatus 1 of this embodiment is demonstrated for every module AC.

(Module A: Initial setting)
FIG. 3 is a flowchart showing the operation of each module of the sleep stage determination device according to the embodiment. In module A, the general nap data acquisition unit 11 acquires nap data of an unspecified number of generally healthy persons (S11). As an index of nap data in this case, typically, an electroencephalogram, a pulse wave, and a heart rate are used, and body temperature, respiration, and the like can be used as a reference. The sleep stages of an unspecified number of people with respect to index values such as these brain waves and pulse waves are acquired as nap data.

  The histogram creation unit 12 creates a histogram of data for each sleep stage of each index (S12). FIG. 4 is a diagram showing a histogram created by module A. As shown in FIG. As illustrated in FIG. 4, the histogram creation unit 12 creates a histogram indicating the frequency of sleep stages S0 to S4 for each of the indicators A1 to An.

  Taking FIG. 5A as an example, the histogram created by the histogram creating unit 12 will be described in more detail. As an example of an index of nap data, an example in which the standard deviation (heart rate fluctuation) of R wave-R wave interval (RRI), which is an R wave time interval at which a pulse wave has a peak on an electrocardiogram, is used as an index Ai of nap data Will be described. It is known that the standard deviation of RRI is closely related to the sleep stage. The histogram of FIG. 5A represents the frequency (frequency) of sleep stages S0 to S4 of unspecified majority persons for each standard deviation value of RRI. The sleep stage for each standard deviation value of the RRI can be determined from the brain wave detection results of the unspecified majority at that time.

  As shown in FIG. 5A, there is a range in which the distributions of the sleep stages S0 to S4 overlap each other with respect to the histogram showing the distribution of the sleep stages with respect to the standard deviation value of the RRI. Here, the initial boundary setting unit 13 divides the sleep range S0 and the sleep step S1 indicated by hatching in the drawing, and initially sets the RRI standard deviation value in which the integrated values of the overlap range are halved. Is set as the boundary value d01 of (S13). The initial boundary setting unit 13 similarly sets initial boundary values d12 to d34 for a range where the sleep stage S1 and the sleep stage S2 overlap to a range where the sleep stage S3 and the sleep stage S4 overlap.

  The change section setting unit 14 sets a range where the sleep stage S0 and the sleep stage S1 overlap as the change section Δ01 (S13). The change section setting unit 14 similarly sets the change sections Δ12 to Δ34 for the range in which the sleep stage S1 and the sleep stage S2 overlap to the range in which the sleep stage S3 and the sleep stage S4 overlap.

  Here, it is considered that a portion having a high degree of overlap between different sleep stage distributions is likely to be in any sleep stage with respect to the RRI standard deviation value, and the reliability to determine that it is in any sleep stage is low. It is done. Therefore, as shown in FIG. 5B, the reliability calculation unit 15 sets the reliability to 0 as the reliability is a value between 0 and 1 and the degree of overlap between the different sleep stage distributions is large. The value is calculated as a close low value, and the reliability is calculated as a high value close to 1 for a portion with a smaller degree of overlap between different sleep stage distributions (S13). In the example of FIG. 5B, the change section Δ01, which is a range where the sleep stage S0 and the sleep stage S1 overlap, is calculated with low reliability.

  Module A similarly repeats the steps S11 to S13 for each of the indices A1 to An, creates a histogram for each of the indices A1 to An, and calculates an initial boundary value, a change interval, and a reliability. These initial values may be sampled based on the age and sex of a generally healthy person.

(Module B: Change of boundary)
Returning to FIG. 3, the user nap data acquisition unit 21 of the module B acquires the nap data of the user who is the subject in a predetermined time of about 15 to 25 minutes (S21). The nap data acquired by the user nap data acquisition unit 21 typically uses an electroencephalogram, a pulse wave, and a heart rate corresponding to the index in the nap data acquired by the module A. Breathing etc. can be used.

  Here, when the user nap data acquisition unit 21 acquires a sufficient number of nap data sample values for each of the indexes A1 to An and updates the initial boundary value set in the module A (S22), the boundary The value / change section acquisition unit 22 acquires the initial boundary value and the change section set in the module A (S23).

  The center-of-gravity calculation unit 23 is within each range divided into two by the initial boundary values d01 to dn-1, n in the change sections Δ01 to Δn−1, n in which the sleep stages S1 to Sn overlap each other. The center of gravity of each range divided by the initial boundary values d01 to dn−1, n is calculated according to the sample values such as the user's pulse wave acquired by the user nap data acquisition unit 21 included (S24).

  FIG. 6 is a diagram illustrating calculation of the center of gravity in the left and right regions of the boundary value in module B. As shown in FIG. 6, for example, in a change section Δi, i + 1 where sleep stages Si to Si + 1 overlap with each other, left Δi, i + 1 to initial which are left regions divided into two by initial boundary values di and i + 1, respectively. The sample value of the index Ak included in the range of the boundary value di, i + 1 is four, and the sample value of the index Ak included in the range of the initial boundary value di, i + 1 to the right Δi, i + 1 that is the right region is six. It is. As shown in FIG. 6, from these sample values, the centroid calculator calculates the left centroid in the range of left Δi, i + 1 to the initial boundary value di, i + 1, and the right centroid in the range of initial boundary value di, i + 1 to right Δi, i + 1. Is calculated.

  The mass ratio calculation unit 24 compares the ratio of the mass of the nap data on the right side of the initial boundary values di, i + 1 with the mass of the nap data on the left side in the change section Δi, i + 1 (S25). For example, there are two sample values of the index Ak included in the range from the left centroid to the initial boundary value di, i + 1, and three sample values of the index Ak included in the range from the initial boundary value di, i + 1 to the right centroid. is there. From now on, the right mass (α): the left mass (1-α) is expressed as “the number of sample values of the index Ak included in the range from the left centroid to the initial boundary value di, i + 1 (two)”: “the initial boundary value di, i + 1 to the number of sample values of the index Ak included in the range of the right centroid (three) ”(0 ≦ α ≦ 1). In this example, α = 2/5 and (1−α) = 3/5.

The boundary value changing unit 25 calculates the changed new boundary value 'di, i + 1 according to the following equation (1) based on α which is the mass ratio calculated by the mass ratio calculating unit 24 (S26). As a result, as shown in FIG. 7, the number of sample values of the index Ak included in the range from the left centroid to the initial boundary value di, i + 1 and the index Ak included in the range from the initial boundary value di, i + 1 to the right centroid. In accordance with the number of sample values, a new boundary value 'di, i + 1 moved so that the side with the larger number of sample values becomes wider in the change section Δi, i + 1 is calculated.
New boundary value 'di, i + 1 = α · “Left center of gravity” + (1−α) · “Right center of gravity” (1)

  The steps S21 to S26 described above are repeated for each initial boundary value d01 to dn-1, n of each index A1 to An every time the user nap data acquisition unit 21 acquires user nap data.

(Module C: Determination of sleep stage)
Returning to FIG. 3, the user nap data acquisition unit 31 of the module C acquires the nap data of the user who is the subject in a predetermined time of about 15 to 25 minutes (S31). The nap data acquired by the user nap data acquisition unit 31 acquires brain waves, pulse waves, heart rate, body temperature, breathing, and the like as indices of nap data, similarly to the nap data acquired by the user nap data acquisition unit 21 of the module B. However, since the boundary value changed for each index is obtained by the module B, for example, the detection of brain waves is omitted, and the type of data that is less than the user nap data acquisition unit 21 of the module B is acquired. it can.

  The reliability comparison unit 33 acquires the reliability of each index A1 to An from the module A and compares the reliability (S32). When the initial boundary value is not updated in the module B (S22) as in the first sleep stage determination, the boundary value / change section acquisition unit 32 sets the initial setting value set in the module A. A period boundary value and a change section are acquired (S33). Further, when the initial boundary value is updated in the module B (S22), the boundary value / change section acquisition unit 32 acquires the updated boundary value and the changed section in the module B (S34). In this case, if it is desired to determine the sleep stage at an early stage, before the initial boundary value is updated in module B, the sleep stage is determined based on the initial boundary value obtained from module A and the change interval. Also good.

  The sleep stage determination unit 34 determines the sleep stage using the index with the highest reliability (S35). FIG. 8 is a diagram illustrating determination of the sleep stage in the module C. As shown in FIG. 8, at the time when the data of each index A1 to An is acquired by the user nap data acquisition unit 31 of the module C, the reliability of the sleep stages S0 to S4 determined by each updated boundary value. Are different. In the example of FIG. 8, the sleep stage determined by the index A1 is the sleep stage S2, and the reliability thereof is 0.5. The sleep stage determined by the index A2 is the sleep stage S2, the reliability is 0.7, the sleep stage determined by the index A3 is the sleep stage S2, and the reliability is 0.3, The sleep stage determined by the indicator An is the sleep stage S3, and its reliability is 0.2.

  The reliability comparison unit 33 compares the reliability of the sleep stage determined by the indices A1 to An at the time as described above, and the sleep stage determination unit 34 determines the sleep stage determined by the index Ak having the highest reliability. Select the sleep stage at that time. In the example of FIG. 8, since the reliability of the sleep stage determined by the index A2 is 0.7, which is the highest reliability, the sleep stage determination unit 34 is a sleep that is the sleep stage determined by the index A2. Stage S2 is selected and set as the sleep stage at that time.

  With the above method, the sleep stage determination unit 34 determines the sleep stage for 20 to 30 minutes as shown in FIG. 9 and stores it as nap data specific to the user who is the subject. Furthermore, as shown in FIG. 10, the sleep stage determination unit 34 calculates the sleep stage allocation ratio for each nap such as the first time and the second time, and stores it as nap data specific to the user who is the subject. This nap data may be organized as user-specific nap data for every day, every week, or every month. Such user-specific nap data can be used, for example, for user health management, sleep quality determination, or sleep disorder diagnosis.

  The driving support ECU 100 determines the quality of the nap of the user who is a driver after a short nap based on the result of the sleep stage determination, and based on the quality of the user's nap and the fatigue level of the user. To control the traveling of the host vehicle. In addition, the driving support ECU 100 makes a notification such as recommending the user to take a break or a nap through a display or a speaker based on the quality of the user's nap and the user's fatigue level.

  According to the present embodiment, the sleep stage determination unit 34 is a sleep stage distribution that is a distribution of biometric features for each individual sleep stage with respect to the biometric features related to the sleep of the individual. Since an individual's sleep stage is determined based on the degree of overlap, it is possible to perform a more accurate determination of the sleep stage as compared to a method of determining a sleep stage using a predetermined threshold.

  Further, according to the present embodiment, the sleep stage distribution is different from the biometric feature quantity in the sleep stage distribution that is the distribution of the biometric feature quantity for each sleep stage with respect to the biometric feature quantity relating to the individual sleep. It is sufficient to determine the sleep stage of the individual from the degree of overlapping, and the driving support ECU 100 provides information related to the return to driving of the individual vehicle based on the sleep stage of the individual determined by the sleep stage determination unit 34. It is possible to perform safe driving after a rest.

  Furthermore, according to the present embodiment, the boundary value changing unit 25 sleeps differently with respect to the biometric feature amount in the sleep stage distribution that is the distribution of the biometric feature amount for each individual sleep stage with respect to the biometric feature amount related to the sleep of the individual. Since the boundary value of the sleep stage is calculated within the range where the stage distributions overlap, and the sleep stage determination unit 34 determines the sleep stage of the individual based on the biometric feature amount and the boundary value, the sleep is performed with a predetermined threshold. Compared with the method for determining the stage, it is possible to determine the sleep stage more accurately.

  In addition, according to the present embodiment, the user nap data acquisition units 21 and 31 acquire the biometric feature amount related to the sleep of the individual, and the initial boundary setting unit 13 determines whether the biometric feature amount related to the sleep of an unspecified number of individuals. In the sleep stage distribution, which is the distribution of biometric features for each specific sleep stage of a large number of individuals, an initial boundary value of the sleep stage is set within a range where different sleep stage distributions overlap with the biometric features, and the sleep stage The determination unit 34 is based on at least one of the biometric feature amount acquired by the user nap data acquisition units 21 and 31 and the initial boundary value set by the initial boundary setting unit 13 and the boundary value calculated by the boundary value changing unit 25. Since the individual's sleep stage is determined, the sleep stage can be determined more accurately based on a large number of unspecified data.

  Further, according to the present embodiment, the boundary value changing unit 25 is a living body in a sleep stage distribution that is a distribution of biometric features for each sleep stage of an unspecified number of individuals with respect to biometric features related to sleep of an unspecified number of individuals. According to the number of sample values of the biometric feature amount acquired by the user nap data acquisition unit 21 included in each range divided into two by the initial boundary value in a range where different sleep stage distributions overlap with the feature amount Then, the boundary value is calculated by correcting the initial boundary value, and the sleep stage determination unit 34 calculates the biometric feature amount acquired by the user nap data acquisition unit 31 and the boundary value change unit 25 by correcting the initial boundary value. In order to determine the sleep stage of the individual based on the boundary value, each time the individual sleeps, the data for determining the sleep degree reflecting the characteristics of the individual is updated, and the heart rate variability associated with the season, physical condition change, etc. Influence It can be eliminated.

  In addition, according to the present embodiment, the center-of-gravity calculation unit 23 is a biometric feature in the sleep stage distribution that is a distribution of biometric features for each sleep stage of the unspecified large number of individuals with respect to the biometric feature related to sleep of the unspecified large number of individuals. In a range in which different sleep stage distributions overlap with each other, each divided by the initial boundary value according to the sample value of the biometric feature amount included in each range divided by the initial boundary value The center of gravity of the range is calculated, and the boundary value changing unit 25 corrects the initial boundary value and calculates the boundary value according to the number of sample values of the biometric feature amount included between each of the center of gravity and the initial boundary value. Therefore, finer correction is possible.

  On the other hand, in the present embodiment, the reliability calculation unit 15 calculates the reliability of the sleep stage with respect to the biometric feature according to the degree of overlap between the sleep stage distributions different from the biometric feature in the sleep stage distribution, Since the sleep stage determination unit 34 determines an individual's sleep stage based on the reliability calculated by the reliability calculation unit 15, the sleep stage determination unit 34 uses the data related to the reliability of the sleep stage with respect to the biometric feature amount. The accuracy can be increased.

  Also, in the histogram of sleep stages, the greater the degree of overlap between sleep stage distributions, the harder it is to determine which sleep stage is, so the reliability is lower, and the lesser the degree of overlap between sleep stage distributions, Since it becomes easy to determine whether it is a sleep stage, reliability becomes high. Therefore, in the present embodiment, the reliability calculation unit 15 calculates the reliability lower as the degree of overlapping of the sleep stage distributions in the sleep stage histogram increases, and the reliability decreases as the degree of the overlap of the sleep stage distributions decreases. By calculating the degree high, it is possible to accurately obtain the reliability of the sleep stage with respect to the biometric feature.

  In the present embodiment, the user nap data acquisition unit 31 acquires a plurality of types of individual biometric features, and the reliability calculation unit 15 determines the biometric features in each sleep stage distribution for each type of biometric feature. In accordance with the degree of overlapping of different sleep stage distributions with respect to the amount, the reliability of the sleep stage for each type of biometric feature is calculated, and the sleep stage determination unit 34 obtains each type acquired by the biometric feature acquisition unit. In order to determine the sleep stage of the individual based on the sleep stage having the highest reliability determined by the reliability calculation unit 15 for the biometric feature quantity, the sleep stage determination based on each type of biometric feature quantity is the most. Those having high reliability will be adopted, and it becomes possible to more accurately determine the sleep stage that further matches the characteristics of the individual.

  In addition, according to the present embodiment, in order to acquire the sleep stage accurately by acquiring the individual's brain wave, using the heart rate data such as the heart rate and heart rate fluctuation of the individual that can acquire the biometric feature amount in a short time. Thus, it is possible to determine the sleep degree level with better immediacy and accuracy.

  It should be noted that the present invention is not limited to the above-described embodiment, and it is needless to say that various modifications can be made without departing from the gist of the present invention.

It is a block diagram which shows the structure of the sleep stage determination apparatus which concerns on embodiment. It is a flowchart which shows the outline of operation | movement of the sleep stage determination apparatus which concerns on embodiment. It is a flowchart which shows the operation | movement in each module of the sleep stage determination apparatus which concerns on embodiment. 6 is a diagram showing a histogram created by module A. FIG. (A) is a histogram created by module A, and (b) is a diagram showing calculation of reliability in module A. It is a figure which shows calculation of the gravity center in the area | region on the right and left of the boundary value in the module B. FIG. It is a figure which shows the boundary value after the change in the module B. FIG. It is a figure which shows determination of the sleep stage in the module C. FIG. It is a figure which shows the sleep stage for every elapsed time determined by the module C. FIG. It is a figure which shows the allocation ratio of the sleep stage for every nap determined by the module C. FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Sleep stage determination apparatus, 11 ... General nap data acquisition part, 12 ... Histogram preparation part, 13 ... Initial boundary setting part, 14 ... Change area setting part, 15 ... Reliability calculation part, 21 ... User nap data acquisition part, 22 ... Boundary value / change interval acquisition unit, 23 ... Center of gravity calculation unit, 24 ... Mass ratio calculation unit, 25 ... Boundary value change unit, 31 ... User nap data acquisition unit, 32 ... Boundary value / change interval acquisition unit, 33 ... Reliability comparison unit, 34 ... sleep stage determination unit, 100 ... travel support ECU, A, B, C ... module.

Claims (20)

  In the sleep stage distribution, which is the distribution of the biometric feature amount for each sleep stage with respect to the biometric feature amount related to the sleep of the individual, the sleep stage of the individual is determined from the degree to which the different sleep stage distributions overlap with the biometric feature amount. The sleep stage determination apparatus provided with the sleep stage determination means to do. In the sleep stage distribution, which is the distribution of the biometric feature amount for each sleep stage with respect to the biometric feature amount related to the sleep of the individual, the sleep stage of the individual is determined from the degree to which the different sleep stage distributions overlap with the biometric feature amount. Sleep stage determination means for
Based on the sleep stage of the individual determined by the sleep stage determination unit, a travel support unit that provides information on driving return to the vehicle of the individual;
A sleep stage determination device comprising: In the sleep stage distribution, which is the distribution of the biometric feature amount for each sleep stage with respect to the biometric feature amount related to the sleep of the individual, the sleep stage boundary value is within a range in which the sleep stage distributions that are different from the biometric feature amount overlap each other. Boundary value calculating means for calculating
Sleep stage determination means for determining the sleep stage of the individual based on the biological feature and the boundary value;
A sleep stage determination device comprising: Biometric feature acquisition means for acquiring the biometric feature related to sleep of the individual;
In the sleep stage distribution, which is the distribution of the biometric feature amount for each sleep stage of the unspecified large number of individuals with respect to the biometric feature amount related to the sleep of the unspecified large number of individuals, the sleep stage distributions that are different from the biometric feature amount overlap each other. An initial boundary value setting means for setting an initial boundary value of the sleep stage within a range;
Further comprising
The sleep stage determination unit includes at least one of the biometric feature amount acquired by the biometric feature amount acquisition unit, the initial boundary value set by the initial boundary value setting unit, and the boundary value calculated by the boundary value calculation unit. The sleep stage determination apparatus according to claim 3, wherein the sleep stage of the individual is determined based on a heel. The boundary value calculation means is different from the biometric feature amount in the sleep stage distribution, which is a distribution of the biometric feature amount for each sleep stage of the unspecified large number of individuals with respect to the biometric feature amount related to sleep of the unspecified large number of individuals. In the range where the sleep stage distributions overlap each other, the initial value according to the number of sample values of the biometric feature amount acquired by the biometric feature amount acquisition unit included in each range divided into two by the initial boundary value Calculate the boundary value by correcting the boundary value,
The sleep stage determination unit is configured to detect the sleep of the individual based on the biometric feature amount acquired by the biometric feature amount acquisition unit and the boundary value calculated by correcting the initial boundary value by the boundary value calculation unit. The sleep stage determination device according to claim 4, wherein the stage is determined.   The boundary value calculation means is different from the biometric feature amount in the sleep stage distribution, which is a distribution of the biometric feature amount for each sleep stage of the unspecified large number of individuals with respect to the biometric feature amount related to sleep of the unspecified large number of individuals. Each range divided into two by the initial boundary value according to the sample value of the biometric feature amount included in each range divided into two by the initial boundary value within a range where the sleep stage distributions overlap. And calculating the boundary value by correcting the initial boundary value according to the number of sample values of the biometric feature amount included between each of the centroids and the initial boundary value. 5. The sleep stage determination device according to 5. In the sleep stage distribution, further comprising a reliability calculation means for calculating the reliability of the sleep stage with respect to the biometric feature according to the degree of overlap between the sleep stage distributions different from each other with respect to the biometric feature,
The sleep stage determination apparatus according to claim 1, wherein the sleep stage determination unit determines the sleep stage of the individual based on the reliability calculated by the reliability calculation unit.   The said reliability calculation means calculates the said reliability low, so that the degree to which the said sleep stage distribution overlaps is large, and calculates the said reliability highly, so that the degree to which the said sleep stage distribution overlaps is small. Sleep stage determination device. The biometric feature acquisition unit acquires a plurality of types of biometric features of the individual,
The reliability calculation means is configured for each type of biometric feature according to the degree of overlap of the different sleep stage distributions with respect to the biometric feature in each of the sleep stage distributions for each type of the biometric feature. Calculating the reliability of the sleep stage,
The sleep stage determination unit is configured to determine the sleep stage of the individual based on the sleep stage with the highest reliability determined by the reliability determination unit with respect to each type of the biometric feature acquired by the biometric feature acquisition unit. The sleep stage determination apparatus according to claim 7 or 8, wherein: The biometric feature is at least one of heartbeat and heartbeat fluctuation of the individual,
The sleep stage determination apparatus according to any one of claims 1 to 9, wherein the sleep stage is acquired based on an electroencephalogram of the individual.   In the sleep stage distribution, which is the distribution of the biometric feature amount for each individual sleep stage with respect to the biometric feature amount related to the sleep of the individual, the sleep stage of the individual is determined based on the degree of overlap of the sleep stage distributions that differ with respect to the biometric feature amount. The sleep stage determination method including the process of determining. In the sleep stage distribution, which is the distribution of the biometric feature amount for each sleep stage of the individual with respect to the biometric feature amount related to the sleep of the individual, the sleep of the individual is determined based on the degree of overlap of the sleep stage distributions that differ with respect to the biometric feature amount. Determining the stage;
Providing information related to the driving return of the individual to the vehicle based on the determined sleep stage of the individual. In the sleep stage distribution that is the distribution of the biometric feature amount for each individual sleep stage with respect to the biometric feature amount related to the sleep of the individual, the sleep stage distributions that are different from each other in the biometric feature amount overlap each other. Calculating a boundary value;
A step of determining a sleep stage of the individual based on the biological feature and the boundary value. Obtaining a biometric feature relating to sleep of the individual;
In the sleep stage distribution, which is the distribution of the biometric feature amount for each sleep stage of the unspecified large number of individuals with respect to the biometric feature amount related to the sleep of the unspecified large number of individuals, the sleep stage distributions that are different from the biometric feature amount overlap each other. Further comprising, within a range, setting an initial boundary value of the sleep stage,
The sleep stage determination method according to claim 13, wherein the sleep stage of the individual is determined based on the acquired biometric feature amount and at least one of the initial boundary value and the boundary value. The sleep stage distributions that are different from the biometric feature amount in the sleep stage distribution that is the distribution of the biometric feature amount for each sleep stage of the unspecified large number of individuals with respect to the biometric feature amount related to sleep of the unspecified large number of individuals overlap. Within the range, the boundary value is calculated by correcting the initial boundary value according to the number of sample values of the acquired biometric feature included in each range divided into two by the initial boundary value,
The sleep stage determination method according to claim 14, wherein the sleep stage of the individual is determined based on the acquired biometric feature amount and the boundary value calculated by correcting the initial boundary value.   The sleep stage distributions that are different from the biometric feature amount in the sleep stage distribution that is the distribution of the biometric feature amount for each sleep stage of the unspecified large number of individuals with respect to the biometric feature amount related to sleep of the unspecified large number of individuals overlap. In the range, according to the sample value of the biometric feature amount included in each range divided into two by the initial boundary value, calculate the center of gravity of each range divided into two by the initial boundary value, The sleep stage determination according to claim 15, wherein the boundary value is calculated by correcting the initial boundary value according to the number of sample values of the biometric feature amount included between each centroid and the initial boundary value. Method. In the sleep stage distribution, the method further includes a step of calculating a reliability of the sleep stage with respect to the biometric feature amount according to a degree to which the different sleep stage distributions overlap with the biometric feature amount,
The sleep stage determination method according to claim 11, wherein the sleep stage of the individual is determined based on the reliability.   The sleep stage determination method according to claim 17, wherein the degree of reliability is calculated to be lower as the degree of overlap between the sleep stage distributions is larger, and the degree of reliability is calculated to be higher as the degree of overlap between the sleep stage distributions is smaller. Obtaining each of the plurality of types of biometric features of the individual,
In each of the sleep stage distributions with respect to each type of the biometric feature amount, the reliability of the sleep stage with respect to each type of the biometric feature amount according to the degree to which the different sleep stage distributions overlap with the biometric feature amount. Respectively,
The sleep stage according to claim 17 or 18, wherein the sleep stage of the individual is determined based on the sleep stage having the highest reliability for each type of the biometric feature quantity acquired by the biometric feature quantity acquisition unit. Judgment method. The biometric feature is at least one of heartbeat and heartbeat fluctuation of the individual,
The sleep stage determination method according to any one of claims 11 to 19, wherein the sleep stage is acquired based on an electroencephalogram of the individual.

Images