JP2009018091A - Doze detector - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a technique capable of highly accurately detecting sleep prediction. <P>SOLUTION: This doze detector includes: first measurement means for measuring a blinking time of a subject; second measurement means for measuring the heart rate fluctuation of the subject; storage means for storing a first reference value, or a reference value of a blinking time, and a second reference value, or the reference value of an HF component of the heart rate fluctuation; first monitor means for monitoring the extension of the blinking time of the subject by a comparison between the maximum value of the blinking time measured by the first measurement means with the first reference value; second monitor means for monitoring a rise in HF component of the heart rate fluctuation of the subject by comparison between the HF component of the heart rate fluctuation measured by the second measurement means with the second reference value; and determination means for determining sleep prediction when the HF component of the heart rate fluctuation is started to increase during appealing the extension of the blinking time. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a technique for detecting dozing.

  Conventionally, many techniques for detecting dozing by monitoring biological information have been proposed.

  Patent Document 1 proposes an apparatus that detects a driver's sleep by measuring a driver's heartbeat with an electrode attached to a handle. Patent Document 2 proposes an apparatus for monitoring a driver's heartbeat with a heartbeat sensor attached to a seat. Patent Documents 3 and 4 propose a method for determining a sign of dozing from the frequency of blinking and the length of time.

Non-patent document 1 proposes a drowsiness prediction method during driving based on an electroencephalogram, a blinking time, the number of blinks, and a heart rate based on an experiment for 12 cases. In Non-Patent Document 2, an international standard is used for sleep evaluation, and 18 subjects are subjected to simultaneous recording of fingertip pulse wave inclination time series, heart rate, respiration, and body surface temperature, and a sign of sleep onset (sleep first) It has been reported that 5-10 minutes before the stage) can be detected.
International Publication No. 2004/089209 Pamphlet JP-A-5-330360 JP 2005-525142 A Japanese Patent Laid-Open No. 9-39603 Numata Nakaho, Kitajima Hiroki, Goi Mihiro, et al., “Study on prediction method of sleepiness when driving a car”, Transactions of the Japan Society of Mechanical Engineers, 63, pp. 101-108, 1997 Yasunori Fujita, Yumi Ogura, Naoki Ochiai, et al., “Development of a method for measuring the onset of sleep onset using fingertip volume pulse wave information”, Ergonomics, 41 (4), pp. 203-212, 2005

  However, the above prior art has the following problems.

  First, devices such as Patent Literatures 1 to 4 do not have an objective evaluation regarding sleepiness and sleep, and medical verification is not sufficient regarding the relevance between biological information such as blinking and heartbeat and sleep. Therefore, doubt remains about the accuracy and reliability of dozing detection. On the other hand, in the researches of Non-Patent Documents 1 and 2, although sleepiness and sleep are evaluated using a medically or internationally recognized evaluation method, an electroencephalogram measuring device is mounted on the vehicle and the driver's Monitoring the electroencephalogram (Non-Patent Document 1) or measuring the body surface temperature by thermography in the vehicle (Non-Patent Document 2) is difficult to put into practical use with the current technology.

  For the purpose of avoiding danger and preventing accidents, it is too late in time to detect the driver's “sleeping” or “sleeping state”. It is desirable to promptly detect a phenomenon before falling asleep, that is, a “predictive of falling asleep”, and give a warning to a driver, control of an automobile, and the like. Furthermore, recent studies have reported that there is a case called microsleep before sleep onset, and such microsleep can be detected from the viewpoint of risk avoidance and accident prevention. Realization of the method is desired. Conventionally, however, there has not been a practical method capable of accurately detecting a sleep onset sign.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a technique capable of accurately detecting a sleep onset sign.

  Although it is relatively easy to monitor the blinking time and is suitable for a simple determination of a sleep onset sign, there is a disadvantage in that the individual difference is large and the accuracy is low. On the other hand, since the heart rate variability analysis can evaluate the autonomic nerve activity, it can be expected to make a dozing determination with higher accuracy than using the blinking time. However, in heart rate variability analysis, it is generally necessary to observe RR interval time series data of 5 minutes or more, and due to its poor time response, it has not been suitable for the determination of sleep onset.

  Therefore, the present inventors made a hypothesis that it is possible to accurately determine a sleep onset sign by simultaneously monitoring two pieces of biological information of blink time and heart rate variability, and conducted an experiment based on a medical viewpoint. As a result, an extension of the blinking time was first observed, followed by the finding that the HF component of heart rate variability began to rise, and then fell asleep and became a sleep state. The configuration of the present invention is based on such knowledge.

  Specifically, the drowsiness detection device according to the present invention is a first measurement means for measuring the blink time of the subject, a second measurement means for measuring the subject's heartbeat fluctuation, and a reference value for the blink time. By comparing the first reference value and the second reference value that is the reference value of the HF component of heart rate variability with the blink time measured by the first measuring means and the first reference value, A comparison between the first monitoring means for monitoring the extension of the subject's blinking time, the HF component (high frequency component) of the heart rate variability measured by the second measuring means, and the second reference value is performed. Second monitoring means for monitoring an increase in the HF component of the heart rate variability, and determination means for determining that the HF component of the heart rate variability is started when an increase in the blinking time is observed, and determining as a sleep onset sign .

  With this configuration, it is possible to detect the sleep onset stage before falling asleep with high accuracy. When a warning sign is detected or a stop signal is output at the time when a sleep symptom is detected, it is possible to prevent a dozing and avoid danger.

  It is preferable that the first reference value is updated according to a measurement value of the subject's blink time measured by the first measurement unit. Moreover, it is also preferable that the second reference value is updated according to the measured value of the subject's heart rate fluctuation measured by the second measuring means.

  There are individual differences in blinking time. In addition, even for the same person, there are day-to-day variations depending on living conditions. It has been clarified from previous reports that heart rate variability varies depending on age, living conditions and the presence or absence of basic diseases such as heart disease. Therefore, when a uniform reference value is used, the determination accuracy is likely to decrease. Therefore, if the reference value is updated as needed using the measurement value of the subject person as in the above configuration, accurate determination can be realized.

  The determination means determines sleep onset when the difference between the HF component of heart rate variability and the second reference value reaches a predetermined value after determining that it is a sign of sleep onset. When it continues for a predetermined period, it is preferable to determine as continuous sleep. Furthermore, it is preferable to output a warning at each stage of sleep onset, sleep on, and continuous sleep.

  The first measuring unit may be capable of measuring a subject's blinking time without contact. For example, the blinking time may be measured using an image obtained by photographing the subject with a camera. Furthermore, it is also preferable to use the image to obtain information such as the inclination of the head (flicker) and the movement of the eyeball, and use it for doze determination.

  The second measuring means may be capable of measuring heartbeat fluctuations of the subject without contact. For example, a method of measuring the heartbeat of the subject person in a non-contact manner using microwaves can be considered.

  However, the present invention does not exclude a configuration for measuring the blinking time or heartbeat fluctuations with a contact-type measurement system. A contact-type measurement system may be used as long as it does not give the subject a great load or discomfort.

  The present invention can be understood as a dozing detection device having at least a part of the above means. The present invention can also be understood as a dozing detection method including at least a part of the above processing or a program for realizing the method. Each of the above means and processes can be combined with each other as much as possible to constitute the present invention.

  For example, the dozing detection method as one aspect of the present invention includes a step of measuring a subject's blink time, a step of measuring the subject's heartbeat fluctuation, and a comparison between the measured blink time and a first reference value. Monitoring the extension of the subject's blinking time, and monitoring the increase in the HF component of the subject's heart rate variability by comparing the measured HF component of the heart rate variability with a second reference value; And a step of determining a sleep onset symptom when an increase in the HF component of heart rate variability is recognized while an extension of the blinking time appears.

  According to the present invention, it is possible to accurately detect a sleep onset sign.

  Exemplary embodiments of the present invention will be described in detail below with reference to the drawings. In the following embodiments, a configuration in which the present invention is applied to a vehicle drowsiness prevention device is illustrated, but the scope of application of the present invention is not limited thereto. For example, the present invention is preferably applicable to scenes where falling asleep may cause danger or accidents, such as operation of trains, airplanes, ships, heavy machinery, or work in factories.

<Device configuration>
FIG. 1 is a block diagram showing a configuration of a dozing detection device according to an embodiment of the present invention.

  This dozing detection device 1 is a device mounted on a vehicle, and is for detecting and preventing a driver's (target person's) dozing. The dozing detection device 1 includes a device body 10, a camera 20, and a heart rate sensor 30 as main components.

  The apparatus main body 10 is an information processing apparatus including a CPU (Central Processing Unit) 11, a storage device 12, an I / F 13 with the camera 20, an I / F 14 with the heart rate sensor 30, an alarm device 15, and the like. The storage device 12 stores various parameters such as various programs executed by the CPU 11 and reference values used in the dozing detection process. The function of the dozing detection device 1 is realized by the CPU 11 executing a program stored in the storage device 12 and controlling the camera 20, the heart rate sensor 30, and the alarm device 15 as necessary. Note that an information processing device such as an ECU or a car navigation system mounted on the vehicle may be used as the device body of the dozing detection device.

  The camera 20 is an imaging device for capturing a driver's video. The image captured by the camera 20 is used in the apparatus main body 10 for measuring the blink time and detecting the tilt of the head. Therefore, the camera 20 is installed at a position where the driver's head and face can be photographed, such as an instrument panel or a handle in front of the driver's seat. As the camera 20, a digital camera using an image sensor such as a CCD or a CMOS is suitable.

  The heart rate sensor 30 is a measurement device for measuring a driver's heart rate variability. The heart rate sensor 30 is a sensor that uses a microwave to measure a driver's heart rate in a non-contact manner. The output signal of the heart rate sensor 30 is taken into the apparatus body 10 and used for heart rate variability analysis.

  As described above, in this embodiment, since non-contact sensors such as a camera and a microwave sensor are used, biological information (blink, heart rate fluctuation) used for detection of falling asleep without giving a load or discomfort to the driver. Obtainable. Therefore, mounting on a vehicle is easy and realistic. However, a contact-type sensor may be used as long as it does not give the driver a great load or uncomfortable feeling. For example, if a contact-type heart rate sensor is provided on a member that originally has contact with the driver, such as a handle, a seat, a headrest, or a seat belt, heart rate variability can be measured without causing a load or a sense of incongruity.

  In the present embodiment, the storage device 12 of the apparatus body 10 corresponds to the storage means of the present invention, the camera 20 corresponds to the first measurement means of the present invention, and the heart rate sensor 30 corresponds to the second measurement means of the present invention. It corresponds to.

<Dozing detection processing>
FIG. 2 is a graph schematically showing the relationship between the driver's condition and blinking time and changes in the HF component of heart rate variability. The horizontal axis of the graph represents time, and the vertical axis represents the blink time and the level (HF value) of the HF component of heart rate variability. The broken line is the change in the blinking time, and the solid line is the change in the HF component of the heart rate variability. Note that the relationship shown in the graph of FIG. 2 is based on data obtained by experiments of the present inventors. Details of this experiment will be described later.

  As can be seen from the graph of FIG. 2, when the driver is in an awake state (and resting), both the blinking time and the HF component of heart rate variability are at low levels. On the other hand, when the driver is in a sleep state, the blink is lost and the HF component of heart rate variability is at a high level. Looking at the transition period from the awake state to the sleep state, it can be seen that an extension of the blinking time appears first, followed by an increase in the HF component of heart rate variability, followed by falling asleep and reaching a sleep state. Note that “sleeping” refers to the start of sleep. The dozing detection device 1 according to the present embodiment performs detection in three stages of “sleep falling predictor”, “sleep falling”, and “sleep state (persistent sleep)” illustrated in FIG. 2.

  Now, the operation of the dozing detection apparatus 1 will be described with reference to the flowchart of FIG. The dozing detection process described here is executed by the CPU 11 of the dozing detection device 1.

(Step S1)
When the process starts, the camera 20 and the heart rate sensor 30 start measurement. A driver's image and a heartbeat signal are taken into the apparatus main body 10 every predetermined time (for example, every several milliseconds). CPU11 calculates a driver | operator's blink time by analyzing the time series image taken in sequentially. A known image processing method can be used for calculating the blink time. For example, there is a method of determining whether the eye is open or closed by detecting eyes, eyelids, pupils, white eyes, etc. by image processing, and obtaining the blinking time by how many images in the closed eye state are continuous. Further, the CPU 11 calculates the level of the HF component of the driver's heart rate variability by analyzing the heartbeat signals that are sequentially captured. Since the method for evaluating the HF component of heart rate variability is known, the description thereof is omitted here.

The calculation result (measurement value) of the blinking time and the HF component of heart rate variability is obtained every predetermined time (for example, 1 second). Using the measurement values that are periodically measured, the processes after step S2 are performed.

(Step S2)
In step S2, the CPU 11 executes a reference value update process. As shown in FIG. 2, the reference value is the HF component value (baseline) of the blinking time and heart rate variability in the driver's resting and awake state.

  There are individual differences in blinking time. In addition, even for the same person, there are day-to-day variations depending on living conditions. It has been clarified from previous reports that heart rate variability varies depending on age, living conditions and the presence or absence of basic diseases such as heart disease. Therefore, when a uniform reference value is used, the determination accuracy is likely to decrease. In particular, the present embodiment aims to accurately detect a sleep onset sign that appears immediately before falling asleep, so it is desirable to eliminate as much as possible the factors that lower the determination accuracy. Therefore, in this embodiment, the reference value is updated as an initialization process before entering the doze determination.

  Specifically, the CPU 11 calculates the reference value (first reference value) of the driver's blink time by averaging the measured values of the blink time obtained during a predetermined initialization period. Similarly, the HF component reference value (second reference value) is obtained by averaging the measured values of the HF component obtained during the initialization period. The reference value calculated here is stored in the storage device 12, and is used for the processing after step S3.

  Thus, by updating the reference value as needed using the measurement value of the driver himself / herself, accurate determination can be made regardless of individual differences or daily fluctuations. The initialization period can be arbitrarily set (for example, it may be set to about several minutes). In the present embodiment, the average of the measurement values is used as a reference value, but other statistical values such as an intermediate value and a mode value may be used.

(Step S3)
When the update process of the reference value is finished, the CPU 11 enters a mode for detecting “a sleep symptom”. First, the CPU 11 monitors the extension of the driver's blink time by comparing the measured value of the blink time with a reference value. The function of step S3 corresponds to the first monitoring means of the present invention.

  For example, the CPU 11 calculates an increase rate of the blinking time every time a measurement value is obtained, and determines that the extension is “extended” when the increase rate exceeds a predetermined threshold value. If extension of the blink time is recognized, the process proceeds to the next step S4.

(Step S4)
In step S4, the CPU 11 monitors the increase in the HF component of the driver's heart rate variability by comparing the measured value of the HF component of heart rate variability with the reference value. The function of step S4 corresponds to the second monitoring means of the present invention.

  For example, the CPU 11 obtains a difference from the reference value every time a measured value of the HF component is obtained, and determines that the HF component has started to rise when the difference exceeds a predetermined threshold value. If the blink time returns to the reference value level during the monitoring operation in step S4, the process returns to the monitoring operation in step S3.

(Step S5)
While the extension of the blink time appears (step S3), when the condition that the start of the increase of the HF component of the heart rate variability is recognized (step S4) is satisfied, the CPU 11 gives the driver a “pregnant sleep sign”. Judge that it appeared. That is, the time point of point A in FIG. 2 is detected as a “sleeping symptom”. In the experiments by the present inventors, it has been confirmed that the point A at which the above conditions are met precedes the actual sleep onset by several seconds to several tens of seconds. Also, as will be described later, when the microsleep phenomenon occurs, the blinking time is prolonged and the HF component is increased. Therefore, the microsleep phenomenon can be detected by using the above determination condition.

(Step S6)
If the CPU 11 determines that it is a “predictive of falling asleep”, the CPU 11 controls the alarm device 15 to output a first-stage warning. By issuing a warning immediately before falling asleep, the driver can be awakened, so accidents and dangers can be avoided in advance.

(Step S7)
After the first stage “pregnant sleep” is detected, the CPU 11 enters a mode for detecting the second stage “sleeping”. As shown in FIG. 2, according to the experiments of the present inventors, the HF component of heart rate variability is in the middle of rising at the time of falling asleep. In this embodiment, in order to capture this state, the rising slope of the HF component is monitored, and when the slope reaches a predetermined threshold, it is determined that the driver has “sleeped”.

  In addition, since the time for which eyes are closed becomes longer after falling asleep and the frequency of blinking decreases, the accuracy of sleep determination may be improved by monitoring the frequency of blinking together with the rising slope of the HF component. Further, since the head tilt starts to occur after falling asleep, the head tilt may be evaluated from the position and tilt of the head in the image, and the information may be used for sleep determination.

  By the way, in the sleep onset detection mode in step S7, when a decrease in the HF component and shortening of the blinking time are recognized, it is considered that the driver is in an awake state, and the operation returns to the sleep onset sign detection mode in step S3 again.

(Step S8)
If the CPU 11 determines “sleeping” in step S <b> 7, the CPU 11 controls the alarm device 15 to output a second-stage warning. At this time, it is preferable to output a warning with a louder sound than in the first stage and to urge the driver to wake up reliably.

(Step S9)
After the second stage “sleeping” is detected, the CPU 11 enters a mode for detecting the third stage “persistent sleep”. As shown in FIG. 2, the HF component of heart rate variability stabilizes at a certain level after continuing to rise for a while after falling asleep. Therefore, in this embodiment, the difference (or rate of increase) between the measured value of the HF component and the reference value is monitored, and when the state where the difference exceeds a predetermined threshold continues for 10 minutes or more, the operation is performed. It is determined that the person has entered “continuous sleep”. Since blinking disappears completely after entering the continuous sleep state, it is also preferable to use the information for determination of continuous sleep. In addition, even if a decrease in the HF component is recognized in this mode, it is considered that the driver has returned to the awake state, and the process returns to the sleep onset detection mode in step S3 or the sleep detection mode in step S7.

(Step S10)
When it determines with "persistent sleep" at step S9, CPU11 controls the alarm device 15 and outputs the warning of a 3rd step.

In steps S6, S8, and S10, not only a warning may be output, but the CPU 11 may send a signal to the ECU to control the operation of the vehicle. For example, for the purpose of avoiding danger, it is assumed that a vehicle horn is sounded, a steering wheel is locked, an engine speed is decreased, or a speed is decreased. In the above flow, step S
3 and S4 are executed in order, but monitoring of the blinking time and monitoring of the HF component of heart rate variability may be executed in parallel by multitask processing.

  According to the drowsiness detection apparatus 1 described above, it is possible to accurately detect the three stages of “asleep sign”, “asleep”, and “a persistent sleep”. In particular, there is an advantage that it is possible to detect a sleep onset symptom and micro sleep, which have been difficult to detect conventionally, with high accuracy. Further, since a non-contact sensor is used, it is excellent in that it can be easily and realistically mounted on a vehicle. Moreover, as will be described below, the decision logic at each stage is based on medical experiments and has high reliability.

<Explanation of experiment>
The outline of the experiment conducted by the present inventors will be described below. The subjects of the experiment were 25 healthy university students.

(1) Experimental environment The experiment was performed in a dark room in a shield room. The subject's posture was set to the sitting position together with driving. Prior to testing, smoking and excessive drinking and exercise were prohibited. The test was stopped when sleep could not be obtained for more than 20 minutes in the dark room in the shield room. Since the first inspection was not used to the environment, data from the second time onward was used.

(2) Sleep, blinking and heart rate variability analysis Sleep was evaluated using polysomnography (PSG) overnight. EEG attachment (C4-A1, C3-A2, O2-A1, O1-A2) and sleep determination were in accordance with international standards. Information obtained from this system is the sleep stage, ECG RR interval, electroencephalogram, electromyogram, and eye movement.

  AntiSleep (manufactured by SmartEye) was used for measurement of blink information. Information obtained from AntiSleep includes head coordinates, head rotation, eyeball coordinates, line-of-sight direction, and eyelid gap.

MemCalc / Win was used for frequency analysis of heart rate variability in the ECG RR interval. The low frequency (LF) and high frequency (HF) components are respectively calculated as frequency integrated values in 0.04 Hz-0.15 Hz and 0.15 Hz-0.4 Hz bands in the power spectrum of the electrocardiogram RR interval time series data.

In this experiment, we investigated the characteristics of the blink information and heart rate variability obtained from AntiSleep around sleep onset by simultaneous measurement of AntiSleep and PSG.

  Paired t-test was used for the test between the two groups, and P <0.05 was considered significant.

(3) Example of Acquired Data by PSG FIG. 4A is an example of brain wave, eye movement, and electrocardiogram data acquired by PSG. In PSG, sleep stage (sleep stage) with 30-second data as 1 unit (epoch)
(It may be determined in a 20-second section.) The Rechtschaffen & Kales classification is widely used for EEG depth classification. Sleep stages 1-4 are collectively referred to as NREM sleep.
The sleep stages 3 and 4 are collectively called slow wave sleep (SWS).

-Awakening (Stage W): Alpha waves of 8-12 Hz account for 50% or more. Rapid eye movement was observed, and the electromyogram was relatively high potential.
-Sleep stage 1 (Stage I): α wave amplitude decrease, 8-12 Hz α wave is less than 50%. EEGs of various frequencies can be seen at low potential. Slow eye movements are observed.
・ Sleep stage 2 (Stage II): EEG, spindles (spindles) appear at the center and the top of the head. The emergence of Kcomplex. Eye movement almost disappeared.
Sleep stage 3 (Stage III): On the electroencephalogram, δ waves having an amplitude of 75 μV or more at 2 Hz or less occupy 20-50% of 1 epoch.
Sleep stage 4 (Stage IV): On the electroencephalogram, δ waves having an amplitude of 75 μV or more at 2 Hz or less occupy 50% or more of 1 epoch.
REM sleep (Stage REM): Electroencephalograms of various frequencies are seen at a relatively low potential. An alpha wave that is 1-2 Hz slower than the awakening may appear. Characteristic rapid eye movement is observed, and the electromyogram has the lowest potential throughout the night.

(4) Results of Heart Rate Variability Analysis in Sleep State and Awake State FIG. 5 shows the results of heart rate variability frequency analysis in the sleep state and the awake state. In this case, LF increased in the awake state. In the awake state, an increase in LF / HF reflecting sympathetic nerve activity was observed. In sleep state, an increase in HF reflecting parasympathetic activity was observed.

(5) Results of heart rate variability analysis in microsleep state and awake state In this experiment, heart rate variability analysis was also performed in the microsleep state and the awake state. FIG. 4 (a) shows sleep stage 1, and FIG. 4 (b) shows brain wave, eye movement, and electrocardiogram RR interval data in the micro sleep state. Micro sleep is `` Manual
of Standardized Terminology, Techniques and Scoring System for Sleep Stage of Human Subjects, Sleep Disorders Atlas Task Force of the American Sleep Disorders Association.EEG arousals: scoring rules and examples, and Blaivas AJ, et al. According to “help assess subjective sleepiness. Sleep Medicine 2007; 8: 156 ~ 159”, it was defined as sleep of 1 epoch for 3 seconds or more and less than 15 seconds.

  As shown in FIG. 6, the LF / HF value is lower in the micro sleep state than in the awake state, similarly to the sleep state. In addition, a slight increase in HF is observed even in the microsleep state.

(6) Analysis of blink information
The blinking time (the time from the eye closing time to the eye opening time) was calculated from the eyelid gap (unit: meters) obtained from AntiSleep. Then, the blinking time in the awake state was compared with the blinking time 10 seconds before falling asleep. As shown in FIG. 7, the blinking time was significantly extended 10 seconds before falling asleep compared to the awake state.

(7) Integration of blink time and heart rate variability analysis index FIG. 8 shows changes over time in the HF component (5-minute value) of the blink time and heart rate variability. In this case, an extension of the blinking time was observed prior to the sleep onset time (390 seconds), and then an increase in the HF value was observed. In other cases, an increase in the blink time and an increase in the HF value were observed at the time of sleep (the transition period from awakening to sleep), and the increase in the blink time preceded the increase in the HF value. The result of this experiment is integrated and generalized in the graph of FIG.

FIG. 1 is a block diagram illustrating a configuration of a dozing detection device. FIG. 2 is a graph schematically showing the relationship between the driver's state and the blinking time and the change in the HF component of heart rate variability. FIG. 3 is a flowchart showing the flow of the dozing detection process. FIG. 4 is an example of electroencephalogram and eye movement data acquired by PSG. FIG. 5 shows heart rate variability frequency analysis results in the sleep state and the awake state. FIG. 6 shows heart rate variability frequency analysis results in the microsleep state and the awake state. FIG. 7 shows a comparison between the awake state and the blinking time 10 seconds before falling asleep. FIG. 8 shows temporal changes of the blinking time and the HF component (5-minute value) of heart rate variability.

Explanation of symbols

1 dozing detection device 10 device body 11 CPU
12 Storage device 13 I / F with camera
14 I / F with heart rate sensor
15 Alarm device 20 Camera 30 Heart rate sensor

Claims (3)

First measuring means for measuring a subject's blink time;
Second measuring means for measuring heart rate variability of the subject;
Storage means for storing a first reference value that is a reference value for blinking time and a second reference value that is a reference value for the HF component of heart rate variability;
First monitoring means for monitoring an extension of the blink time of the subject by comparing the blink time measured by the first measurement means with the first reference value;
Second monitoring means for monitoring an increase in the HF component of the subject's heart rate variability by comparing the HF component of heart rate variability measured by the second measuring means with the second reference value;
A determination means for determining a sleep onset symptom when an increase in the HF component of the heart rate variability is recognized while an extension of the blinking time appears;
A dozing detection device comprising: The dozing detection apparatus according to claim 1, wherein the first reference value is updated according to a measurement value of the blink time of the subject measured by the first measurement unit. The dozing detection apparatus according to claim 1 or 2, wherein the second reference value is updated in accordance with a measured value of heartbeat variability of the subject measured by the second measuring unit.

Images