JP2001061820A - Sleeping state judging method and sleeping state judging device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately detect a sleeping state by calculating a trend curve for showing an increase/decrease tendency of a time change in time series of living body information such as the heartbeat number, and calculating the trend curve by using a window function of preset time when judging a sleeping state on the basis of the curve. SOLUTION: A measuring part 1 is arranged to measure a living body information value such as the heartbeat number per unit time as a living body information measured value, and a trend curve for showing a tendency of a time change in time series of the living body information value is calculated by a trend curve calculating part 3. An increase in the living body information value in each measured time is calculated by an increase calculating part 4 from the living body information value and the trend curve to calculate a fluctuation degree of the living body information value in a specific time before each measured time by a fluctuation degree calculating part 5. An appearance probability of respective sleeping states is calculated by a calculating part from a combination of the increase and the fluctuation degree of the living body information value, and a state of adopting a maximum numeric value among the appearance probability of the respective sleeping states is estimated as a sleeping state of the corresponding measured time by an estimating part 8.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sleep state judging method for detecting a change in the sleep state of a living body on the basis of relatively easily obtained information on the activity of the living body such as a heart rate and a pulse rate. The present invention relates to a sleep state determination device.

[0002]

2. Description of the Related Art In general, a heartbeat signal is constantly changing due to the state of the human being, for example, sleep / wake, rest / exercise, change in posture, eating, emotion, mental relaxation and stress. Regarding sleep, the human sleep state is not uniform throughout the night, and a cycle between the non-REM sleep phase and the REM sleep phase appears periodically several times, and the cycle is about 100 minutes (80 to 120 minutes). It has been known for some time. In each cycle, the sleep state gradually shifts from light sleep to deep sleep in the non-REM sleep period, after the deep sleep state is maintained, the light sleep state is returned again, and then the change in the appearance of the REM sleep period is general It is a target. Further, the change in the sleep depth during the non-REM sleep period in each cycle is relative, and the sleep depth becomes shallower for each cycle from falling asleep to awakening. Therefore, if a change in the state of sleep as described above is detected, appropriate stimuli can be given to promote sleep onset according to the state, to induce a deeper sleep state, and to wake up comfortably. it can.

Therefore, various attempts have been made to detect changes in the state of sleep, and for example, a polysomnogram including brain waves, eye movements, and myoelectricity may be used. However, this device is large in scale and can be used only in a place equipped with measurement equipment such as a laboratory or a hospital, and is not suitable for daily use such as health equipment. Therefore,
It is desired to accurately detect a change in the state of sleep by means that is an alternative to polysomnography. In order to respond to such a demand, it has been considered that attention is paid to a heart rate and a pulse rate, and a change in the state of a human being is detected from these changes.

[0004] That is, in night sleep, it is known that the heart rate per unit time decreases with falling asleep, shows a minimum value in the morning, and increases as the awakening time approaches. In the non-REM sleep period, the heart rate per unit time is relatively stable, but in the REM sleep period, the autonomic nervous system activity is disturbed, and the heart rate remarkably fluctuates. It is known that Further, even during the non-REM sleep period, when the sleep depth is shallow, the heart rate per unit time is relatively large, and when the sleep depth is deep, the heart rate per unit time is relatively small and very stable. FIG. 17 (a) shows an example of transitions of four sleep states overnight, including an awake state, a REM sleep period, a light sleep state during a non-REM sleep period, and a deep sleep state during a non-REM sleep period, and FIG. b) shows an example of the change of the heart rate at this time. In FIG. 17, “WAKE” is awake, “REM” is REM sleep, and “non-RE”.
“M” indicates a light sleep state during the non-REM sleep period, and “SWS” indicates a deep sleep state during the non-REM sleep period.

Various configurations for detecting a sleep state based on such knowledge have been conventionally provided. For example, as disclosed in Japanese Patent Application Laid-Open No. 63-205592, an index of change in pulse rate is used as an index. There is one that detects the REM sleep period. Also, as disclosed in Japanese Patent Application Laid-Open No. 3-41926, it is possible to accurately detect a change in sleep state by using a sleep index incorporating the increasing and decreasing tendency of the pulse rate and respiratory rate and the temporal variation. There is something.

[0006]

However, in the above-mentioned conventional apparatus, falling asleep is discriminated by using a sleep index incorporating the increasing and decreasing tendency of the pulse rate and respiratory rate and temporal fluctuation, and the REM sleep period is determined. Because it was detected, there is considerable individual difference in the matching rate with the actual change in sleep state,
The accuracy was considerably lower than that of polysomnography, and there was a problem that a deep sleep state and a light sleep state could not be distinguished.

An object of the present invention is to solve the above problems, and an object of the present invention is to provide a sleep state judging method and a sleep state judging device capable of accurately detecting a sleep state.

[0008]

According to a first aspect of the present invention, there is provided a sleep state determining method which measures biological information such as a heart rate or a pulse rate and indicates a tendency of a time series of the biological information to increase or decrease over time. In a sleep state determination method for calculating a trend curve and determining a sleep state based on the trend curve, the trend curve is calculated using a window function of a preset time.

According to a second aspect of the present invention, in the first aspect, when calculating the trend curve using the window function, the length of time of the window function is set to 40 to 70 minutes. is there.

According to a third aspect of the present invention, there is provided a measuring unit for measuring a biological information value such as a heart rate or a pulse rate per unit time to obtain a biological information measurement value at each time.
A measurement time setting unit 2 for setting a measurement time; a trend curve calculation unit 3 for calculating a trend curve set so as to represent a tendency of a time series change of the biological information value; An increment calculator 4 that calculates the degree of increase of the biological information value at each measurement time, a variation degree calculator 5 that calculates the degree of variation of the biological information value within a fixed time before each of the measurement times, A sleep state appearance probability calculation unit 7 that calculates an appearance probability of each sleep state based on a preset sleep state appearance probability density distribution table 6 from a combination of the calculated increment of the biological information value and the degree of dispersion;
A sleep state estimating unit 8 for estimating the state that takes the largest numerical value among the appearance probabilities of each sleep state calculated by the sleep state appearance probability calculation unit 7 as the sleep state at the measurement time. It is a feature.

According to a fourth aspect of the present invention, there is provided a sleep state judging device which measures a biological information value such as a heart rate or a pulse rate per unit time and obtains a biological information measurement value at each time.
A measurement time setting unit 2 for setting a measurement time; a trend curve calculation unit 3 for calculating a trend curve set so as to represent a tendency of a time series change of the biological information value; An increment calculator 4 that calculates the degree of increase of the biological information value at each measurement time, a variation degree calculator 5 that calculates the degree of variation of the biological information value within a fixed time before each of the measurement times, A sleep state appearance probability calculation unit 7 that calculates an appearance probability of each sleep state based on a preset sleep state appearance probability density distribution table 6 from a combination of the calculated increment of the biological information value and the degree of dispersion;
A sleep state appearance ratio calculation unit 10 that calculates an appearance ratio of each sleep state based on a preset appearance ratio table of each sleep state from an elapsed time from the start of the operation; Sleep state appearance ratio calculation unit 11 that calculates each sleep state appearance ratio by multiplying each sleep state appearance ratio
And a sleep state estimating unit 8 for estimating the state having the largest numerical value among the sleep state appearance ratios as the sleep state at the measurement time.

According to a fifth aspect of the present invention, there is provided a sleep state judging device which measures a biological information value such as a heart rate or a pulse rate per unit time and obtains a biological information measurement value at each time.
A measurement time setting unit 2 for setting a measurement time; a pre-trend curve calculation unit 12 for calculating a pre-trend curve set so as to represent a tendency of a temporal change of the biological information value in a time series traveling direction; A post-trend curve calculating unit 13 that calculates a trend curve after setting so as to represent the tendency of the time-series change of the biological information value in the time series in the reverse direction, and which is larger at each measurement time among the pre-trend curve and the post-trend curve. A trend curve calculating unit 14 for calculating a trend curve by using the two values, an increment calculating unit 4 for calculating the degree of increase of the biological information value at each measurement time from the biological information value and the trend curve, and A variation degree calculation unit 5 that calculates the variation degree of the biological information value within a certain time period before the time, and a combination of the increment of the calculated biological information value and the variation degree, The sleep state appearance probability calculation unit 7 that calculates the appearance probability of each sleep state based on the set sleep state appearance probability density distribution table 6 and the largest numerical value among the appearance probabilities of the respective sleep states are taken. And a sleep state estimating unit 8 for estimating the state as the sleep state at the measurement time.

According to a sixth aspect of the present invention, there is provided a sleep state judging device which measures a biological information value such as a heart rate or a pulse rate per unit time and obtains a biological information measurement value at each time.
A measurement time setting unit 2 for setting a measurement time; a pre-trend curve calculation unit 12 for calculating a pre-trend curve set so as to represent a tendency of a temporal change of the biological information value in a time series traveling direction; A post-trend curve calculating unit 13 that sets a trend curve of the time series of the biological information value in the reverse direction and then calculates a trend curve, and a large value at each measurement time among the pre-trend curve and the post-trend curve. A trend curve calculating unit 14 for calculating a trend curve by using the two values, an increment calculating unit 4 for calculating the degree of increase of the biological information value at each measurement time from the biological information value and the trend curve, and A variation degree calculation unit 5 that calculates the variation degree of the biological information value within a certain time period before the time, and a combination of the increment of the calculated biological information value and the variation degree, Sleep state appearance probability calculation unit 7 that calculates the appearance probability of each sleep state based on the sleep state appearance probability density distribution table 6 set in advance, and a sleep state that is set in advance from the elapsed time from the start of operation. A sleep state appearance ratio calculation unit 10 that calculates the appearance ratio of each sleep state based on the appearance ratio table 9, and calculates each sleep state appearance ratio by multiplying each sleep state appearance probability and each sleep state appearance ratio. Sleep state appearance ratio calculation unit 11 and a sleep state estimation unit 8 for estimating the state with the largest numerical value among the sleep state appearance ratios as the sleep state at the measurement time. It is assumed that.

According to a seventh aspect of the present invention, in the third to sixth aspects, after the sleep state estimating section 8 estimates that the sleep state is estimated to fall asleep, a deep sleep is performed for a preset time. It is characterized by including a sleep determination adjusting unit 15 for correcting the estimation result so as not to be in a state.

According to the invention of claim 8, in the claims 3 to 7, the estimated result output from the sleep state estimating unit 8 is estimated to fall asleep and then the degree of variation is less than a preset threshold value. When a certain state is continuously maintained for a preset time, a sleep determination adjusting unit 16 that corrects an estimation result so as to be in a deep sleep state is provided.

[0016]

Embodiments of the present invention will be described below. In the following embodiments, the heart rate is used as a biological information value that is an index of the activity state of a living body.
It goes without saying that the same processing can be performed for the pulse rate.

FIG. 1 shows an example of a basic configuration of a control unit of an apparatus for carrying out the first and second aspects of the present invention. The measurement of the heart rate can be performed using a heart rate sensor that detects the beat of the heart using electrodes or radio waves, and the output from the heart rate sensor is transmitted by wireless transmission or wired transmission to this device. Is transmitted to In this device, the heart rate signal transmitted from the heart rate sensor is received and input to the measuring unit 1. After obtaining a pulse-like signal by performing waveform shaping, the measuring unit 1 counts the number of pulses per unit time, and counts the counted value as the heart rate H per unit time.
(T). The unit time for counting the number of pulses is, for example, one minute. The heart rate per unit time H (t) output from the measurement unit 1 is input to the trend curve calculation unit 3 and the sleep state estimation unit 8.

The trend curve calculation unit 3 calculates a trend curve set so as to represent an increasing / decreasing tendency in a time series of a biological information value by using a window function of a preset time (h). It is. The window function is a function used to limit the section to be analyzed when performing analysis on certain continuous data. In the present invention, the time of the window function is determined by the REM sleep in which the heart rate tends to increase. In order to extract the characteristics of the period, it is preferable that the period is longer than the duration of the REM sleep period. Conversely, in order to better reflect the trend of the heart rate, the time of the window function should be shorter. Considering these two conditions together, it is considered that the window function time is optimally between 40 minutes and 70 minutes, especially about 60 minutes.

Subsequently, the trend curve calculator 3 operates as shown in the flowchart of FIG. 2, and defines the trend curve FT (t) from the input heart rate H (t) per unit time as follows. I do.

When t = 1 FT (1) = H (1) When 1 <t <h When H (t) ≧ FT (t−1) FT (t) = FT (t−1) H ( t) <FT (t-1) FT (t) = H (t) When h ≦ t FT (t) = min (H (t−h + 1)... H (t)) The trend curve FT (t) is
For example, as shown in FIG. In FIG. 3, the heart rate is represented by a thin black line, and the trend curve is represented by a thick black line.

The trend curve FT (t) calculated by the trend curve calculating section 3 is output from the trend curve calculating section 3 and input to the sleep state estimating section 8. Then, the sleep state estimating unit 8 uses the trend curve FT (t) and the heart rate H (t) to calculate the 4th state of the awake state, the REM sleep period, the light sleep state of the non-REM sleep period, and the deep sleep state of the non-REM sleep period. It can be determined by estimating which kind of sleep state it is. Trend curve FT
(T) and the method of estimating which of the four sleep states it is using heart rate H (t) are not particularly limited, but are the same as in the case of FIGS. In addition, the increase in the heart rate is calculated from the difference between the trend curve FT (t) and the heart rate H (t), and the difference between the one-minute heart rate and the next one-minute heart rate is sequentially calculated to obtain a variation. Calculate the degree,
The probability of appearance of four types of sleep states is obtained from the combination of the increase in heart rate and the degree of variability, and the appearance ratio of four types of sleep states is obtained based on the elapsed time since entering the bed. The sleep state with the largest (appearance ratio) can be determined by estimating the sleep state at that time.

FIG. 4 shows an example of the basic configuration of the control unit of the device according to the third aspect of the present invention. The output from the heart rate sensor is transmitted to the apparatus by wireless transmission or wired transmission, and the heart rate signal transmitted from the heart rate sensor is input to the measuring unit 1. After obtaining a pulse-like signal by performing waveform shaping, the measuring unit 1 counts the number of pulses per unit time, and counts the counted value as the heart rate H per unit time.
(T). The unit time for counting the number of pulses is, for example, one minute. The heart rate H (t) per unit time output from the measurement unit 1 is input to the trend curve calculation unit 3, the increase calculation unit 4, and the variation calculation unit 5.

The trend curve calculating section 3 calculates a trend curve set so as to represent an increasing / decreasing tendency of a biological information value with time in a time series. That is, the trend curve calculation unit 3 operates as shown in the flowchart of FIG. 5, and defines the trend curve FT (t) from the input heart rate H (t) per unit time as in the following equation.

When t = 1 FT (1) = H (1) When t> 1 When H (t) ≧ FT (t−1) FT (t) = FT (t−1) H (t) <FT (t-1) FT (t) = H (t) The trend curve FT (t) thus obtained is
For example, as shown in FIG. In FIG. 6, the heart rate is indicated by a thin black line, and the trend curve is indicated by a thick black line.

The trend curve FT (t) calculated and output by the trend curve calculating section 3 is input to an increment calculating section 4 for calculating an increment which is the degree of increase of the biological information value at each measurement time. You. The increment calculation unit 4 calculates an increment ZOU (t) which is an index indicating the degree of increase of the heart rate H (t) from the trend curve FT (t). That is, the increment ZOU (t) is defined as follows using the heart rate H (t) and the trend curve FT (t).

ZOU (t) = H (t) -FT (t) The thus obtained increment ZOU (t) is, for example, as shown in FIG. FIG. 7 shows the heart rate H (t) of FIG.
And the trend of the overnight increment obtained from the trend curve FT (t).

The variability calculating section 5 calculates the variability of the heart rate H (t) input from the measuring section 1. First, the measured heart rate H (t)
And the absolute value of the difference between the heart rate H (t-1) and the heart rate H (t-1) measured one minute before that (referred to as difference SABUN (t)). The formula for calculating the difference SABUN (t) is defined as follows.

If abs (H (t) -H (t-1)) <6, SABUN (t) = abs (H (t) -H (t-1)) abs (H (t) -H (t) -1)) ≧ 6 SABUN (t) = 6 Further, the degree of variation BA is calculated using the difference SABUN (t).
RA (t) is calculated as in the following equation.

When t = 1 BARA (1) = SABUN (1) When t = 2 BARA (1) = SABUN (1) + SABUN (2) When t = 3 BARA (1) = SABUN (1) + SABUN (2)
+ SABUN (3) When t = 4 BARA (1) = SABUN (1) + SABUN (2)
+ SABUN (3) + SABUN (4) When t ≧ 5 BARA (t) = SABUN (t−4) + SABUN
(T-3) + SABUN (t-2) + SABUN (t-
1) + SABUN (t) The variation degree BARA (t) obtained as described above is, for example, as shown in FIG.

As described above, the increment ZOU (t) calculated and output by the increment calculating section 4 and the variation BARA calculated and output by the variation calculating section 5
(T) is a sleep state appearance probability calculation unit 7 that calculates the appearance probability of each sleep state based on the sleep state appearance probability density distribution table 6.
Is input to The sleep state appearance probability calculation unit 7 calculates a sleep state appearance for each sleep state from a combination of the increment ZOU (t) output from the increment calculation unit 4 and the variation BARA (t) output from the variation degree calculation unit 5. Based on the probability density distribution table 6, the appearance probabilities of the four types of sleep states of the awake state, the REM sleep period, the light sleep state of the non-REM sleep period, and the deep sleep state of the non-REM sleep period are calculated. ing.

The sleep state appearance probability density distribution table 6 is created based on night sleep state data for 89 nights (about 42,000 minutes in time), for example. That is, the increment and the degree of variation of the heart rate are calculated every minute and minute of the data of about 42,000 minutes.
The sleep state is determined from the brain wave data every minute.
Then, by analyzing the relationship between the combination of the one-minute heart rate increment and the degree of variation and the type of sleep state from these data, the one-minute sleep state is a light sleep state in the non-REM sleep period. It is possible to calculate the probability, the probability of the REM sleep period, the probability of the deep sleep state in the non-REM sleep period, and the probability of the awake state, respectively. The sleep state appearance probability density distribution table 6 summarizes these probabilities, and as shown in FIG. 9 as an example, the light sleep state (non-REM) in the non-REM sleep period and the REM sleep period (R
EM), a non-REM sleep deep sleep state (SWS), and an awake state (WAKE) for each sleep state. Appearance probability density distribution tables 6a, 6b,
6c and 6d are two-dimensional tables so that the appearance probability of the sleep state can be obtained from the combination of the increment and the degree of variation. For example, as shown in FIG. 9, when the increment is “3” and the variation is “5”, the appearance probability of the light sleep state (non-REM) in the non-REM sleep period is 68.5.
%become. And the sleep state appearance probability calculation unit 7
The appearance probability for each sleep state calculated and output in
It is input to the sleep state estimation unit 8.

The sleep state estimating unit 8 compares the appearance probabilities of the four sleep states output from the sleep state appearance probability calculation unit 7 and estimates the sleep state having the largest value as the sleep state at that time. As described above, it is possible to estimate any one of the four sleep states of the awake state, the REM sleep period, the light sleep state of the non-REM sleep period, and the deep sleep state of the non-REM sleep period.

Next, FIG. 10 shows an example of the basic configuration of the control unit of the apparatus according to the fourth aspect of the present invention. The measuring unit 1, the trend curve calculating unit 3, the increment calculating unit 4, the variation calculating unit 5, and the sleep state appearance probability calculating unit 7 are the same as those in FIG. The appearance probability for each sleep state calculated and output by the sleep state appearance probability calculation unit 7 is input to the sleep state appearance ratio calculation unit 11.

The sleep state appearance ratio calculation unit 10 calculates the sleep state in the awake state, the REM sleep period, and the non-REM sleep period based on the appearance ratio table 9 of each sleep state based on the elapsed time from the start of the sleep operation. , Non-REM, deep sleep, 4
The appearance ratio of each type of sleep state is calculated. The sleep state appearance ratio table 9 analyzes the relationship between the elapsed time from the entrance and the type of sleep state based on the data of the sleep state at night, as described above, to thereby obtain four types of sleep states. It is created by calculating the appearance ratio, and the sleep state when a predetermined time has elapsed from the start of sleep is awake state, REM sleep period, light sleep state in non-REM sleep period, deep sleep in non-REM sleep period FIG. 11 shows the ratio of the awake state (WAK state).
E), REM sleep period (REM), non-REM light sleep state (non-REM), non-REM deep sleep state (SWS), appearance ratio of each of the four sleep states Table 9
Is shown as a graph. Sleep state appearance ratio calculation unit 6
The sleep state appearance ratio calculated and output is input to the sleep state appearance ratio calculation unit 11.

The sleep state appearance ratio calculation unit 11 calculates the appearance probability of each sleep state output from the sleep state appearance probability calculation unit 7 and the sleep state appearance ratio output from the sleep state appearance ratio calculation unit 10. To calculate the appearance ratio of each sleep state. The sleep state appearance ratio calculated and output by the sleep state appearance ratio calculation unit 11 is input to the sleep state estimation unit 8.

The sleep state estimating unit 8 compares the appearance ratios for each sleep state output from the sleep state appearance ratio calculation unit 11 and estimates the state showing the largest value as the sleep state at that time. Yes, it is possible to estimate which of the four sleep states, awake state, REM sleep period, light sleep state during non-REM sleep period, and deep sleep state during non-REM sleep period.

Next, FIG. 12 shows an example of the basic configuration of the control unit of the device according to the fifth aspect of the present invention. Measurement unit 1
Is the same as in the case of FIG. Then, the heart rate H (t) per unit time output from the measurement unit 1 is output to the front trend curve calculation unit 12, the rear trend curve calculation unit 13, the increment calculation unit 4, and the variation calculation unit 5.

The pre-trend curve calculating section 12 calculates a pre-trend curve FT (t) which is set so as to represent an increasing / decreasing tendency of the biological information value in the time series in the time direction. The operation of calculating the previous trend curve FT (t) is the same as the operation of the trend curve calculation unit 3 in FIG. Then, the front trend curve FT (t) output from the front trend curve calculation unit 12 is input to the trend curve calculation unit 14.

The post-trend curve calculating section 13 calculates a post-trend curve RT (t) that is set so as to represent an increasing / decreasing tendency in a time series in the reverse direction of the time series of the biological information value. That is, the post-trend curve calculation unit 13
Operates as shown in the flowchart of FIG. 13, and defines the post-trend curve RT (t) from the input heart rate H (t) per unit time as in the following equation.

Ds: Final measurement time When t = ds RT (ds) = H (ds) When t <ds H (t) ≧ RT (t + 1) RT (t) = RT (t + 1) H ( t) <RT (t + 1) RT (t) = H (t) Then, the post-trend curve RT (t) calculated and output by the post-trend curve calculation unit 13 is calculated by the trend curve calculation unit 1
4 is input.

The trend curve calculating section 14 outputs the previous trend curve FT output from the previous trend curve calculating section 12.
(T) and the later trend curve RT (t) output from the later trend curve calculation unit 13 adopting the one that takes a larger value at each measurement time.
(T) is calculated. The trend curve TRE (t) calculated and output by the trend curve calculation unit 14 in this manner is input to the increment calculation unit 4 which calculates the degree of increase of the biological information value at each measurement time.

In the increment calculating section 4, the trend curve TRE
An increment ZOU (t), which is an index indicating the degree of increase of the heart rate H (t) from (t), is calculated. That is, the incremental ZO
U (t) is expressed as heart rate H (t) and trend curve TRE (t).
Is defined as follows.

ZOU (t) = H (t) -TRE (t) The variation calculator 5 is the same as that in FIG.

The increment ZOU (t) calculated and output by the increment calculator 4 and the variance BARA (t) calculated and output by the variability calculator 5 are represented by a sleep state appearance probability density distribution table. 6 is input to a sleep state appearance probability calculation unit 7 that calculates the appearance probability of each sleep state, and the appearance probability for each sleep state is calculated. The appearance probability for each sleep state is input to the sleep state estimation unit 8, and the awake state,
It is estimated which of the four types of sleep states, the REM sleep phase, the light sleep state during the non-REM sleep phase, and the deep sleep state during the non-REM sleep phase. The sleep state appearance probability calculating section 7 and the sleep state estimating section 8 are the same as those in FIG.

FIG. 14 shows an example of the basic configuration of the control unit of the device according to the sixth aspect of the present invention. The measurement unit 1, the front trend curve calculation unit 12, the rear trend curve calculation unit 13, the trend curve calculation unit 14, the increment calculation unit 4, the variation degree calculation unit 5, and the sleep state appearance probability calculation unit 7
It is the same as that of FIG. The sleep state appearance ratio calculation unit 10, sleep state appearance ratio calculation unit 11, and sleep state estimation unit 8 are the same as those in FIG.

The appearance probability of each sleep state calculated and output by the sleep state appearance probability calculation unit 7 and the appearance ratio of each sleep state calculated and output by the sleep state appearance ratio calculation unit 10 are respectively The appearance ratio is input to the state appearance ratio calculation unit 11, and the appearance ratio of each sleep state is calculated. The appearance ratio of each sleep state is input to the sleep state estimation unit 8, and is any one of four sleep states: awake state, REM sleep period, light sleep state during non-REM sleep period, and deep sleep state during non-REM sleep period. Estimate.

FIG. 15 shows an example of the basic configuration of the control unit of the device according to the seventh aspect of the present invention. The sections from the measuring section 1 to the sleep state estimating section 8 are the same as those in FIG.

The sleep state estimation result estimated by the sleep state estimation unit 8 is output from the sleep state estimation unit 8 and input to the sleep determination adjustment unit 15. The sleep determination adjustment unit 15 estimates that the user has fallen asleep based on the input estimation result of the sleep state, and then estimates that the user is not determined to be in the deep sleep state for a preset time (for example, 10 minutes). The result is corrected.

Next, FIG. 16 shows an example of the basic configuration of the control unit of the device according to the eighth aspect of the present invention. The sections from the measuring section 1 to the sleep state estimating section 8 are the same as those in FIG.

The sleep state estimation result estimated by the sleep state estimation unit 8 is output from the sleep state estimation unit 8 and input to the sleep determination adjustment unit 16. The sleep determination adjustment unit 16 estimates that the user falls asleep based on the input estimation result of the sleep state, and then continuously changes the state in which the degree of variation is less than the preset threshold for a preset period of time. If the duration is maintained for a while, the estimation result is corrected so as to determine that the subject is in a deep sleep state.

[0051]

As described above, in the sleep state determination method according to the first aspect of the present invention, biological information such as a heart rate or a pulse rate is measured, and a time series of the biological information is represented by an increasing or decreasing tendency. In the sleep state determination method of calculating a trend curve and determining a sleep state based on the trend curve, the trend curve is calculated using a window function of a preset time, so that the sleep state can be accurately calculated. In addition, it can be detected in real time.

According to the second aspect of the present invention, when calculating the trend curve using the window function, the length of time of the window function is set to 40 to 70 minutes, so that the sleep state can be detected with high accuracy. Things.

According to a third aspect of the present invention, there is provided a sleep state determining apparatus which measures a biological information value such as a heart rate or a pulse rate per unit time and obtains a biological information measured value at each time; A measurement time setting unit that sets a trend curve calculation unit that calculates a trend curve set to represent the tendency of the time change of the biological information value over time, and a measurement curve at each measurement time from the biological information value and the trend curve. An increment calculation unit that calculates the degree of increase of the biological information value, and a variation degree calculation unit that calculates the degree of variation of the biological information value within a certain time period before each of the measurement times, A sleep state appearance probability calculation unit that calculates an appearance probability of each sleep state based on a preset sleep state appearance probability density distribution table from a combination of the increment and the degree of variation, and a sleep state appearance probability calculation unit. A sleep state estimating unit that estimates the state that takes the largest numerical value among the appearance probabilities of the sleep states as the sleep state at the measurement time. The state appearance probability can be calculated, and the sleep state can be detected accurately and in real time.

According to a fourth aspect of the present invention, there is provided a sleep state judging device for measuring a biological information value such as a heart rate or a pulse rate per unit time to obtain a biological information measured value at each time, A measurement time setting unit that sets a trend curve calculation unit that calculates a trend curve set to represent the tendency of the time change of the biological information value over time, and a measurement curve at each measurement time from the biological information value and the trend curve. An increment calculation unit that calculates the degree of increase of the biological information value, and a variation degree calculation unit that calculates the degree of variation of the biological information value within a certain time period before each of the measurement times, From the combination of the increment and the degree of variation, from a sleep state appearance probability calculation unit that calculates the appearance probability of each sleep state based on a preset sleep state appearance probability density distribution table, and from the elapsed time from the start of operation, The sleep state appearance ratio calculation unit that calculates the appearance ratio of each sleep state based on the appearance ratio table of each sleep state that has been set, and multiplies each sleep state appearance probability by each sleep state appearance ratio Since it has a sleep state appearance ratio calculation unit that calculates the sleep state appearance ratio, and a sleep state estimation unit that estimates the state that takes the largest numerical value among the sleep state appearance ratios as the sleep state at the measurement time. In addition to the increase degree and the variation degree of the biological information value, the appearance ratio of each sleep state can be calculated from the sleep state appearance ratio table, and the concept of time lapse can be added, and the sleep state can be accurately calculated. Moreover, it can be detected in real time.

According to a fifth aspect of the present invention, there is provided a sleep state judging device for measuring a biological information value such as a heart rate or a pulse rate per unit time to obtain a biological information measured value at each time, A measurement time setting unit that sets a pre-trend curve calculation unit that calculates a pre-trend curve that is set to represent a tendency of a time change in the traveling direction of the time series of the biometric information value, The post-trend curve calculation unit that calculates the trend curve after setting to represent the tendency of the time change in the reverse direction, and adopts the larger value at each measurement time among the front trend curve and the rear trend curve. A trend curve calculation unit for calculating a trend curve, an increment calculation unit for calculating the degree of increase of the biological information value at each measurement time from the biological information value and the trend curve, and a constant before the measurement time. A variation degree calculating unit for calculating the degree of dispersion of the biological information value within between, a combination of incremental and variation degree of the calculated biological information value,
A sleep state appearance probability calculation unit that calculates the appearance probability of each sleep state based on a preset sleep state appearance probability density distribution table, and the state that takes the largest numerical value among the appearance probabilities of each sleep state. Since it has a sleep state estimation unit that estimates the sleep state at the measurement time, the degree of increase of the biological information value can be more accurately determined based on the trend curve calculated from the front trend curve and the rear trend curve. It is possible to calculate the appearance probability of each sleep state from the increase degree and the variation degree of the biological information value, and to accurately calculate the sleep state,
Moreover, it can be detected in real time.

According to a sixth aspect of the present invention, there is provided a sleep state judging device for measuring a biological information value such as a heart rate or a pulse rate per unit time to obtain a biological information measured value at each time; A measurement time setting unit that sets a pre-trend curve calculation unit that calculates a pre-trend curve set so as to represent a tendency of a temporal change in the traveling direction of the time series of the biological information value, and a time series of the biological information value A post-trend curve calculation unit that calculates a trend curve after setting it to represent the tendency of the time change in the reverse direction, and adopts the larger value at each measurement time among the above-mentioned front trend curve and the above-mentioned rear trend curve. A trend curve calculation unit for calculating a trend curve, an increment calculation unit for calculating the degree of increase of the biological information value at each measurement time from the biological information value and the trend curve, and a constant before the measurement time. A variation degree calculating unit for calculating the degree of dispersion of the biological information value within between, a combination of incremental and variation degree of the calculated biological information value,
A sleep state appearance probability calculation unit that calculates the appearance probability of each sleep state based on a preset sleep state appearance probability density distribution table, and a predetermined sleep state appearance ratio based on the elapsed time from the start of operation. A sleep state appearance ratio calculating unit that calculates an appearance ratio of each sleep state based on a table, and a sleep state appearance ratio that calculates each sleep state appearance ratio by multiplying each sleep state appearance probability by each sleep state appearance ratio. Since it has a ratio calculation unit and a sleep state estimation unit that estimates the state with the largest numerical value among the sleep state appearance ratios as the sleep state at the measurement time, it is calculated from the front trend curve and the rear trend curve The degree of increase in the biological information value can be calculated as a value that more accurately reflects the state of the living body based on the trend curve that is obtained. In addition, the appearance ratio of each sleep state can be calculated from the sleep state appearance ratio table, and the concept of elapse of time can be added, so that the sleep state can be detected accurately and in real time. .

According to a seventh aspect of the present invention, the estimated result output from the sleep state estimating unit is estimated such that the user falls asleep, and then the estimated result is set so as not to be in a deep sleep state for a preset time. Since the sleep state adjusting unit for correcting the sleep state is provided, the sleep state can be accurately detected.

According to the eighth aspect of the present invention, the state in which the degree of variation is less than a preset threshold is continuously determined after the sleep state estimating unit is estimated to fall asleep. Since the sleep determination adjustment unit that corrects the estimation result so as to be in the deep sleep state when the sleep state is continued for the set time is provided, the sleep state can be accurately detected.

[Brief description of the drawings]

FIG. 1 is a block diagram of a control unit according to an embodiment of the present invention.

FIG. 2 is a flowchart showing the operation of the above trend curve calculation unit.

FIG. 3 is a graph showing a change in heart rate over night and a trend curve corresponding thereto.

FIG. 4 is a block diagram of a control unit according to an embodiment of the third invention.

FIG. 5 is a flowchart showing the operation of the above trend curve calculation unit.

FIG. 6 is a graph showing a change in heart rate over night and a trend curve corresponding thereto.

FIG. 7 is a graph showing a transition of an increase in a heart rate in the same night as above.

FIG. 8 is a graph showing the degree of variation in the heart rate over the same night.

FIG. 9 is a diagram showing a sleep state appearance probability density distribution table according to the first embodiment;

FIG. 10 is a block diagram of a control unit according to an embodiment of the present invention.

FIG. 11 is a graph of a sleep state appearance ratio table according to the embodiment.

FIG. 12 is a block diagram of a control unit according to an example of the embodiment of the present invention;

FIG. 13 is a flowchart showing the operation of the above post-trend curve calculation unit.

FIG. 14 is a block diagram of a control unit according to an example of the embodiment of the present invention;

FIG. 15 is a block diagram of a control unit according to an embodiment of the present invention;

FIG. 16 is a block diagram of a control unit according to an embodiment of the invention;

17A is a graph showing an example of a transition of a sleep state overnight, and FIG. 17B is a graph showing an example of a transition of a heart rate overnight.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Measurement part 2 Measurement time setting part 3 Trend curve calculation part 4 Increment calculation part 5 Variation degree calculation part 6 Sleep state appearance probability density distribution table 7 Sleep state appearance probability calculation part 8 Sleep state estimation part 9 Sleep state appearance ratio table 10 Sleep State appearance ratio calculation unit 11 Sleep state appearance ratio calculation unit 12 Front trend curve calculation unit 13 Back trend curve calculation unit 14 Trend curve calculation unit 15 Sleep judgment adjustment unit 16 Sleep judgment adjustment unit

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Norio Nakano 1048 Kadoma Kadoma, Kadoma City, Osaka Pref. Inside Matsushita Electric Works, Ltd. (72) Inventor Kei Hagiwara 1048 Kazuma Kadoma, Osaka Prefecture F-term in Matsushita Electric Works Co., Ltd. (reference) 4C017 AA02 AA10 AC16 BC11 4C038 PP05 PQ00 PS00

Claims (8)

[Claims] 1. A method of measuring biological information such as a heart rate or a pulse rate, calculating a trend curve representing an increasing / decreasing tendency of a time series change of the biological information, and determining a sleep state based on the trend curve. A sleep state determination method, wherein a trend curve is calculated using a window function for a preset time. 2. The sleep state determination method according to claim 1, wherein, when calculating the trend curve using the window function, the length of time of the window function is set to 40 minutes to 70 minutes. 3. A measuring unit that measures a biological information value such as a heart rate or a pulse rate per unit time and obtains a biological information measurement value at each time, a measurement time setting unit that sets a measurement time, and a biological information value. A trend curve calculating unit that calculates a trend curve set to represent the tendency of the time change of the time series, and an increment calculating unit that calculates the degree of increase of the biological information value at each measurement time from the biological information value and the trend curve. And a variation calculation unit that calculates the variation of the biological information value within a certain time period before each of the measurement times, and a combination of the increment and the variation of the calculated biological information value, which is set in advance. The sleep state appearance probability calculation unit that calculates the appearance probability of each sleep state based on the sleep state appearance probability density distribution table, and the largest among the sleep state appearance probabilities calculated by the sleep state appearance probability calculation unit A sleep state estimating device comprising: a sleep state estimating unit for estimating a state in which a numerical value is taken as a sleep state at the measurement time. 4. A measuring unit that measures a biological information value such as a heart rate or a pulse rate per unit time to obtain a biological information measurement value at each time, a measurement time setting unit that sets a measurement time, and a biological information value. A trend curve calculating unit that calculates a trend curve set to represent the tendency of the time change of the time series, and an increment calculating unit that calculates the degree of increase of the biological information value at each measurement time from the biological information value and the trend curve. And a variation calculation unit that calculates the variation of the biological information value within a certain time period before each of the measurement times, and a combination of the increment and the variation of the calculated biological information value, which is set in advance. A sleep state appearance probability calculation unit that calculates the appearance probability of each sleep state based on the sleep state appearance probability density distribution table, and an appearance ratio table of each sleep state that is set in advance from the elapsed time from the start of the operation. When A sleep state appearance ratio calculation unit that calculates the appearance ratio of each sleep state, and a sleep state appearance ratio calculation unit that calculates each sleep state appearance ratio by multiplying each sleep state appearance probability and each sleep state appearance ratio. A sleep state estimating unit for estimating a state having the largest numerical value among the sleep state appearance ratios as a sleep state at the measurement time. 5. A measuring unit for measuring a biological information value such as a heart rate or a pulse rate per unit time to obtain a biological information measurement value at each time, a measurement time setting unit for setting a measurement time, and the biological information A pre-trend curve calculation unit that calculates a pre-trend curve set to represent a time-varying trend in a time-series traveling direction of a value, and a time-series trend of the biometric information value in a retrograde direction. A post-trend curve calculating unit that calculates a post-trend curve after setting, a trend curve calculating unit that calculates a trend curve by adopting a larger value at each measurement time among the pre-trend curve and the post-trend curve, An increment calculation unit that calculates the degree of increase of the biological information value at each measurement time from the biological information value and the trend curve, and calculates the degree of variation of the biological information value within a certain time before each of the measurement times. A sleep state calculating unit that calculates an appearance probability of each sleep state based on a preset sleep state appearance probability density distribution table from a combination of the calculated increment of the biological information value and the degree of dispersion, A sleep state comprising: an appearance probability calculating unit; and a sleep state estimating unit for estimating a state having the largest numerical value among the appearance probabilities of the respective sleep states as a sleep state at the measurement time. Judgment device. 6. A measuring unit for measuring a biological information value such as a heart rate or a pulse rate per unit time to obtain a biological information measurement value at each time, a measurement time setting unit for setting a measurement time, and the biological information A pre-trend curve calculation unit that calculates a pre-trend curve set to represent a time-varying trend in a time-series traveling direction of a value, and a time-series trend of the biometric information value in a retrograde direction. A post-trend curve calculation unit that calculates a post-trend curve after setting, a trend curve calculation unit that calculates a trend curve by adopting a larger value at each measurement time among the pre-trend curve and the post-trend curve, An increment calculation unit that calculates the degree of increase of the biological information value at each measurement time from the biological information value and the trend curve, and calculates the degree of variation of the biological information value within a certain time before each of the measurement times. A sleep state calculating unit that calculates an appearance probability of each sleep state based on a preset sleep state appearance probability density distribution table from a combination of the calculated increment of the biological information value and the degree of dispersion, An appearance probability calculation unit, a sleep state appearance ratio calculation unit that calculates an appearance ratio of each sleep state based on a preset sleep state appearance ratio table from an elapsed time from the start of the operation, and the sleep state A sleep state appearance ratio calculation unit that calculates each sleep state appearance ratio by multiplying the appearance probability and each of the sleep state appearance ratios, and the state that takes the largest numerical value among the sleep state appearance ratios as the measurement time A sleep state determination device, comprising: a sleep state estimation unit that estimates a sleep state. 7. A sleep determination adjustment unit that corrects the estimation result output from the sleep state estimation unit after estimating that the user has fallen asleep, so as not to be in a deep sleep state for a preset time. The sleep state judging device according to any one of claims 3 to 6, further comprising: 8. A state in which the degree of variability is less than a preset threshold for the estimation result output from the sleep state estimating unit, and the state where the degree of variation is less than a preset threshold continuously continues for a preset time. The sleep state determination device according to any one of claims 3 to 7, further comprising a sleep determination adjustment unit that corrects an estimation result so as to be in a deep sleep state.