JPH05212003A - Sleep detector - Google Patents

Abstract

PURPOSE:To detect the existence of a person with no contact and to judge whether the person is asleep or awaken only when the person is existent. CONSTITUTION:An infrared sensor 1 detects infrared rays radiated from a man M and obtains an output at a level corresponding to the amount of infrared rays made incident. The output of the infrared sensor 1 is detected by a level detection part 3. When the output level of the infrared sensor 1 detected by the level detection part 3 is higher than a prescribed level, a noise component is removed from the output of the infrared sensor 1 by a noise removal part 4. Further, the output of the noise removal part 4 is differentiated by a differential arithmetic part 5 so as to calculate the strength of body action. A fuzzy arithmetic part 6 judges the asleep/awaken state by applying a prescribed membership function to the strength of body action.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sleep onset detecting device capable of contactlessly detecting whether or not a person has fallen asleep.

[0002]

2. Description of the Related Art Conventionally, as a sleep detection device of this type, a pyroelectric infrared sensor, which is a differential sensor, is used to detect a change in the amount of infrared rays caused by a person's body movement, and a pyroelectric infrared sensor. The output value of is sampled in a fixed cycle, and the moving average of the sampling values for each predetermined period (the average value of the sampling values for each predetermined period) and the moving range (the difference between the maximum and minimum values of the sampling values for each predetermined period) A method of determining whether or not a person has fallen asleep based on is considered.

[0003]

In the above configuration, since the body movement of the person is detected by the pyroelectric infrared sensor, the state in which no person is moving and the state in which there is no person are the same. Will be detected as. That is, when there is no person, there is a problem that it may be determined that the body movement has stopped and may be mistakenly recognized as falling asleep.

An object of the present invention is to solve the above problems, and to provide a sleep onset detecting device capable of detecting the presence or absence of a person and determining sleep onset based on the body movement only when the person is present. It is the one we are trying to provide.

[0005]

In order to achieve the above object, the present invention provides an infrared sensor which is arranged at a position where infrared rays emitted from a human body can be detected and which can obtain an output of a level according to the amount of incident infrared rays. And a level detection unit for detecting the output level of the infrared sensor, and a predetermined membership function for fluctuations in the output level of the infrared sensor when the output level of the infrared sensor detected by the level detection unit is equal to or higher than a predetermined level. And a fuzzy calculation unit that determines a sleep / wake state by applying the above.

[0006]

According to the above construction, an infrared sensor capable of obtaining an output of a level corresponding to the amount of infrared rays radiated from the human body is provided, and when the output level of the infrared sensor is above a predetermined level, the sleep / wake state Since the presence of a person is determined based on the output level of the infrared sensor, the sleep / wake state can be determined only when the person is present. It is possible to prevent the person from erroneously recognizing that he / she has fallen asleep due to the determination that the body movement has stopped.

Further, since a predetermined membership function is applied to the fluctuation of the output level of the infrared sensor to judge the sleep / wake state, the position of the infrared sensor with respect to the human body, the physical condition, the individual difference, etc. Therefore, it is possible to flexibly deal with a factor that changes the output level of the infrared sensor and obtain a correct determination result.

[0008]

EXAMPLE As shown in FIG. 1, an infrared sensor 1 is arranged at a position where infrared rays emitted from a person M who is about to sleep can be detected in a non-contact manner. The infrared sensor 1 can obtain an output at a level according to the amount of incident infrared rays, and for example, a photodiode or a thermopile is used.

The output of the infrared sensor 1 is detected by a person M in the detection area.
If there is no person, the level becomes low, and if there is a person M in the detection area, the level increases. Therefore, if the output level of the infrared sensor 1 is detected, the presence or absence of the person M in the detection area can be determined. Further, when the person M is in the detection area of the infrared sensor 1, the amount of infrared rays incident on the infrared sensor 1 is
The change rate of the output level of the infrared sensor 1 frequently changes if the body movement occurs frequently. Since body motion always occurs within a certain period of time during awakening, and almost no body motion occurs during sleep, it is instantaneous. Therefore, a differential value (difference value) is calculated for the output level of the infrared sensor 1, and the differential value is calculated. By determining how the body movement occurs based on the change in the value, it is possible to determine whether the user is awake or sleeping.

Since the output of the infrared sensor 1 is analog data, it is converted into digital data in the A / D converter 2 for later processing. This digital data is input to the level determination unit 3 and compared with a predetermined threshold value. If it is equal to or greater than the threshold value, it is determined that the person M is in the detection area of the infrared sensor 1, and the noise removal unit 4 is activated. on the other hand,
When the digital data input to the level determination unit 3 is smaller than the threshold value, it is determined that the person M is not in the detection area of the infrared sensor 1 and the noise removal unit 4 is prohibited from processing the digital data.

If the digital data is equal to or higher than the threshold value set in the level determination unit 3, the digital data is input to the noise removal unit 4 and sampled at a constant period (several seconds), and a predetermined number of sampling values are started from each sampling value. For each predetermined period (several minutes) in which is obtained, the total variation amount, which is the sum of the differences between the adjacent sampling values in time series, is obtained. The sampling value input to the noise removing unit 4 is x
If the time series is ₁ , x ₂ , x ₃ , ..., a predetermined number of sampling values (x ₁ , ..., x _n , x ₂ , ..., x _{n +1} , x ₃ ,
,, x _{n + 2} , ...) are set respectively, and adjacent sampling values x _i-1 , in each period, are set.
The sum of the differences of x _i (= Σ | x _i-1 −x _i |) is obtained as the total fluctuation amount. The total fluctuation amount corresponds to the total fluctuation range of the output level of the infrared sensor 1 within the above period. When there is little body movement, the difference between the sampling values that are adjacent in time series is small, so the total fluctuation amount is small, and when there are many body movements irregularly, the difference between the sampling values that are adjacent in time series is large. The total fluctuation amount becomes a large value. That is, the degree of body movement can be evaluated based on the total variation.

Further, the noise removing unit 3 obtains the total variation and then obtains the difference value between the adjacent total variations in time series.
That is, if the total variation is y _i , y _{i + 1} −y _i is obtained and output. For example, for the sampling value x _k , | x _{k + n-1} −x _{k + n} | − | x _k−1 −x _k | In this way, the total amount of fluctuation for the predetermined period is calculated,
Further, since the difference value of the total fluctuation amount is obtained, it is possible to remove the influence of instantaneous body movement or disturbance. As a result of such processing, a body movement signal obtained by removing the effects of momentary body movement and disturbance from the output of the infrared sensor 1 as shown in FIG. 3A is obtained. The body movement signal has a large value when the fluctuation of the output level of the infrared sensor 1 is large.

The body movement signal output from the noise removing section 4 is input to the differential calculating section 5. The differential operation unit 5 obtains a discrete differential value of the body movement signal. That is, the differential operation section 5 outputs the differential values in a time series by sequentially obtaining the difference values of a pair of adjacent body movement signals. After all,
Since the output value of the differential operation unit 5 becomes the second-order differential value (second-order difference value) of the total fluctuation amount with respect to time, it represents the change in the change rate of the output level of the infrared sensor 1, that is, the strength of body movement. .. The output value of the differential operation section 5 is input to the fuzzy operation section 6.

The fuzzy calculation section 6 determines a wakefulness state or a sleep state based on the data of the three periods set for the output value of the differential calculation section 5 and the previous output result.
That is, the average value of differential values (input value I ₁ ) for the past 20 minutes as shown in FIG. 3B and the average value (input value I ₂ ) of differential values for the past 5 minutes as shown in FIG. 3C. Then, the average value (input value I ₃ ) of the differential values from 30 minutes before to 5 minutes before in the past as shown in FIG. 3D is used as the three types of data to the fuzzy calculation unit 6. In the fuzzy operation unit 6, as shown in FIG.
The three types of input values I _{1 to} I ₃ and the previous output value O held in the register 11 are input to the _three rule units 12 _{1 to} 12 ₃ in which the membership functions are set. A membership function based on an experiment is set in each of the rule units 12 _{1 to} 12 ₃ , and the rule unit 12 ₁ has a sleep state degree, the rule unit 12 ₂ has a REM sleep degree, and the rule unit 12 ₃ has an awake state. Of the output value Z ₁ ~
It can be obtained as Z ₃ (FIG. 5A-).
(See (c)). Here, the membership function can be set by setting the degree of sleep state to p and the degree of awake state to (1-p), but the membership functions for the sleep state and the awake state may be set individually. .. Also, input values I _{1 to} I
Although ₃ has an average value of a differential value in each period, by setting the membership functions such as using all of the differential values in each period rule unit 12 ₁ to 12 _3, sleep, REM sleep Alternatively, each degree of awake state may be obtained.

By the way, the three rule parts 12 _{1 to} 12 ₃
The output values Z _{1 to} Z ₃ of the
Only the maximum value of the two output values Z _{1 to} Z ₃ is selected and output from the rule selection unit 13. The output value of the rule selection unit 13 is input to the holding unit 14, and in the holding unit 14, the output value selected by the rule selection unit 13 among the output values Z _{1 to} Z ₃ of the _three rule units 12 _{1 to} 12 ₃ is output. Based on this, the transition of sleep / wake states is detected. That is, the awake state as shown in FIG. 3E to the sleep state, the sleep state as shown in FIG. 3F to the REM sleep state, the sleep state as shown in FIG. 3G to the awake state, and FIG. The following four transition patterns from the REM sleep state to the sleep state are detected. Further, the holding unit 14 determines which of the sleep state, the REM sleep state, and the awake state the current sleep / wake state is based on the above-described four types of transition patterns, and outputs the determination result. (See FIG. 3 (i) and FIG. 5 (d)). The determination result is held until the transition pattern changes next time. The determination result is input to the register 11 and used for the next determination.

The portion surrounded by the alternate long and short dash line in FIG. 1 is composed of a microcomputer, and the operation is summarized as shown in FIG. That is, first, the output of the infrared sensor 1 is taken in (step S1), and when the level determining section 3 determines that an output level equal to or higher than a predetermined level is obtained, the noise removing section 4 removes the noise component (step S1). S2). Further, the strength of the body movement is obtained by the differentiation operation in the differentiating unit 5, and the average value of the body movement strength for the past 20 minutes, the past 5 minutes, and the past 30 minutes to the past 5 minutes is obtained. (Step S3). These three types of values are the input signals to the fuzzy operation unit 6, and the output values Z of the rule units 12 _{1 to} 12 ₃ having the maximum function value by applying the three types of membership functions.
Select ₁ to Z ₃ (step S4), and than is determining the state of sleep-wake on the basis of the output value Z ₁ to Z _3, which is selected (step S5).

According to the above construction, the person M by detecting the output level of the infrared sensor 1 by the level determination section 3
The presence / absence of the person can be determined, and the sleep state, the REM sleep state, and the awake state are determined only when the person M is present. Therefore, it is possible to prevent an erroneous determination that he / she fell asleep when the person M was not present. .. In addition, since the fuzzy operation unit 6 is provided to set the membership function for determining the sleep state, the REM sleep state, and the awake state, and the state in which the function value of the membership function is maximum is determined as the current state, It is possible to flexibly deal with the factors that change the output level of the infrared sensor 1, such as the position of the person M with respect to the sensor 1, the daily physical condition, individual differences, and the three states of sleep state, REM sleep state, and awake state. It is possible to prevent erroneous determination of.

[0018]

As described above, the present invention is provided with an infrared sensor capable of obtaining an output of a level corresponding to the amount of infrared rays radiated from the human body, and sleeps when the output level of the infrared sensor is equal to or higher than a predetermined level. Since the awakening state is determined, the presence or absence of a person can be determined based on the output level of the infrared sensor, and the sleep / wake state can be determined only when the person is present. The present invention has an effect that it is possible to prevent a person from erroneously recognizing that he / she has fallen asleep when it is determined that the body movement has stopped when there is no person.

Further, since a predetermined membership function is applied to the fluctuation of the output level of the infrared sensor to judge the sleep / wake state, the position of the infrared sensor with respect to the human body, the physical condition, the individual difference, etc. The advantage is that it is possible to flexibly deal with a factor that changes the output level of the infrared sensor and obtain a correct determination result.

[Brief description of drawings]

FIG. 1 is a block circuit diagram showing an embodiment.

FIG. 2 is an operation explanatory view showing the processing procedure of the embodiment.

FIG. 3 is an operation explanatory diagram showing signals of respective parts of the embodiment.

FIG. 4 is a block circuit diagram showing a fuzzy arithmetic unit used in the embodiment.

FIG. 5 is an operation explanatory diagram of the fuzzy calculation unit in the embodiment.

[Explanation of symbols]

 1 infrared sensor 2 A / D converter 3 level determination unit 4 noise eliminator 5 differential calculator 6 fuzzy calculator

Claims (1)

[Claims] 1. An infrared sensor arranged at a position where infrared rays emitted from a human body can be detected, and an output of a level corresponding to the amount of incident infrared rays is obtained, a level detection section for detecting the output level of the infrared sensor, and a level. A fuzzy calculation unit that determines a sleep / wake state by applying a predetermined membership function to a change in the output level of the infrared sensor when the output level of the infrared sensor detected by the detection unit is equal to or higher than a predetermined level; A sleep onset detection device comprising: