JPWO2018105459A1 - Sleepiness estimation device - Google Patents

Abstract

A drowsiness calculation apparatus capable of calculating drowsiness with high accuracy is provided. A sleepiness estimation apparatus 1 calculates a cumulative average heart rate obtained by dividing a cumulative total of heart rate acquired by an I / F 7 by time after a user's activity state has changed by a sleepiness reference heart rate calculation unit 4. Based on the accumulated average heart rate, a drowsiness reference heart rate that is a reference for whether or not the user feels drowsiness is calculated. Then, the sleepiness calculation unit 5 estimates the user's current sleepiness based on the sleepiness reference heart rate and the current heart rate.

Description

  The present invention relates to a sleepiness estimation apparatus that estimates user sleepiness.

  For example, in order to detect a drowsy driving of a moving body such as a vehicle, it has been proposed to detect a drowsiness state from the heart rate of the driver (see, for example, Patent Documents 1 and 2).

  In Patent Document 1, the average value of the RRI (RR Interval) value of the heartbeat detected by the heartbeat sensor at awakening and the integral multiple of the integrated value of the RRI value exceeding the average value are set as threshold values, and the RRI value exceeding the average value is integrated. In the case where the threshold value is exceeded, it is described that it is determined that the patient is dozing.

  In Patent Document 2, heart rate interval data obtained by time-series RRI is obtained based on a heartbeat waveform of a subject (user) in a resting state measured by a sensor. Next, frequency analysis is performed on the heartbeat interval data, and a frequency analysis result of heartbeat fluctuation including power spectral density (PSD) and autonomic nerve total power (TP) with respect to a frequency at a certain time is obtained. Next, the frequency analysis result was set as the initial state, and based on the estimated drowsiness position and the estimated awakening position included in the initial state, the origin of the drowsiness position and the origin of the awakening position were set. Determine the drowsiness scale. And it is described that the sleepiness of the subject is determined based on the sleepiness scale.

Japanese Patent No. 3252586 JP 2014-12042 A

  In the case of the method described in Patent Document 1, data at the time of initial awakening is used as a reference. However, the initial state varies depending on the person, and the average value of the RRI value and the like cannot be appropriately set depending on the awakening level of the initial state. Therefore, there has been a problem that it cannot cope with the difference in the initial state depending on the person.

  The method described in Patent Document 2 has the following problems. The correspondence between heart rate information and sleepiness scale varies depending on various factors. Even based on a resting state, there are changes due to the day and changes due to time zones. For example, it is known that the time when sleepiness becomes strong is from 3 to 4 in the early morning, and from 3 to 4 in the afternoon. Therefore, it is difficult to estimate sleepiness with high accuracy only from heartbeat information. In addition, for example, driving is a kind of stress load state, and the heart rate baseline is different even in a road state (urban area or highway). Furthermore, since the state at the beginning of measurement is not constant, it has not been possible to cope with changes such as sleepiness from the beginning and sleepiness at the beginning.

  Therefore, in view of the above-described problem, an object of the present invention is to provide a sleepiness estimation apparatus that can estimate sleepiness with high accuracy, for example.

  In order to solve the above-mentioned problem, the invention according to claim 1 is characterized in that an acquisition means for acquiring a user's heart rate and a total of the heart rate acquired by the acquisition means after the activity state of the user has changed. A cumulative average heart rate calculating means for calculating a cumulative average heart rate divided by time, and a drowsiness for calculating a drowsiness reference heart rate as a reference whether or not the user feels drowsy based on the cumulative average heart rate A sleepiness estimation apparatus comprising: a reference heart rate calculation means; and sleepiness estimation means for estimating the user's current sleepiness based on the sleepiness reference heart rate and the current heart rate.

  The invention according to claim 7 is a drowsiness estimation method of a drowsiness estimation device that estimates a user's current drowsiness, an acquisition step of acquiring a user's heart rate, and an activity state of the user is changed. A cumulative average heart rate calculation step of calculating a cumulative average heart rate obtained by dividing a cumulative total of the heart rate acquired by the acquisition step by time, and whether or not the user feels sleepy based on the cumulative average heart rate A drowsiness reference heart rate calculating step for calculating a drowsiness reference heart rate as a reference, and a drowsiness estimation step for estimating the current sleepiness of the user based on the drowsiness reference heart rate and the current heart rate. This is a method for estimating sleepiness.

  The invention described in claim 8 is a sleepiness estimation program characterized by causing a computer to execute the sleepiness estimation method according to claim 7.

It is a schematic block diagram of the drowsiness calculation apparatus concerning 1st Example of this invention. It is the table | surface which showed the example of discrimination | determination of the active state in the state discrimination | determination part shown by FIG. It is the table | surface which showed the example of the parameter table shown by FIG. It is another example of composition which calculates a sleepiness standard heart rate. It is the graph which compared the sleepiness estimated with the sleepiness reference | standard heart rate calculation type | formula of the sleepiness reference | standard of a present Example in a certain subject. It is the graph which compared the sleepiness estimated with the sleepiness reference | standard heart rate calculation type | formula of the sleepiness reference | standard of a present Example in a certain subject. It is the graph which compared the sleepiness estimated with the sleepiness reference | standard heart rate calculation type | formula of the sleepiness reference | standard of a present Example in a certain subject. It is explanatory drawing of the example of a display of the drowsiness display part shown by FIG. It is a flowchart of operation | movement of the drowsiness calculation apparatus shown by FIG. It is a schematic block diagram of the drowsiness calculation apparatus concerning the 3rd Example of this invention. It is a schematic block diagram of the device shown by FIG.

  Hereinafter, a sleepiness estimation apparatus according to an embodiment of the present invention will be described. In the sleepiness estimation apparatus according to the embodiment of the present invention, the cumulative average heart rate calculation unit calculates a cumulative average heart rate obtained by dividing the total of the heart rate acquired by the acquisition unit after the change in the user activity state by time. Then, the drowsiness reference heart rate calculation means calculates a drowsiness reference heart rate that serves as a reference for whether or not the user feels drowsiness based on the cumulative average heart rate. Then, the sleepiness estimation means estimates the user's current sleepiness based on the sleepiness reference heart rate and the current heart rate. By doing so, drowsiness can be estimated based on the accumulated heart rate, so the drowsiness reference heart rate can be calculated in a short time, and drowsiness can be estimated with higher accuracy.

  The accumulated average heart rate calculating means may calculate the accumulated average heart rate after a predetermined time has elapsed after the activity state has changed. By doing so, the influence of other activity states before the change can be reduced, and the calculation accuracy of the sleepiness reference heart rate can be improved.

  Further, the apparatus further comprises a minimum heart rate determination unit that determines a minimum heart rate that is a minimum value of the heart rate per predetermined time acquired by the acquisition unit, and the drowsiness reference heart rate calculation unit includes the cumulative average heart rate and the minimum heart rate. May be calculated as the drowsiness reference heart rate. By doing so, the mechanism for calculating the drowsiness reference heart rate is simplified, so that it is possible to reduce processing time and development costs in the apparatus and the like.

  The drowsiness reference heart rate calculating means may calculate the cumulative average heart rate as the drowsiness reference heart rate. By doing so, it is possible to further simplify the mechanism for calculating the drowsiness reference heart rate, and to reduce the processing time and the development cost in the apparatus or the like. In addition, by setting the cumulative average heart rate as the drowsiness reference heart rate, it is possible to capture a sign of feeling drowsiness, and to take a measure such as prompting a break before recognizing drowsiness.

  Further, the cumulative average heart rate calculating means may calculate the cumulative average heart rate after the user has changed to the active state of driving the mobile body. In this way, when driving a moving body such as a vehicle, the drowsiness reference heart rate can be calculated in a short time, and drowsiness can be estimated with high accuracy.

  A state monitoring apparatus according to an embodiment of the present invention includes the drowsiness estimation apparatus described above and a measurement unit that measures a user's heart rate. By doing in this way, when monitoring a user's state, a user's sleepiness can be estimated with high precision. Therefore, the monitoring accuracy can be improved.

  In addition, the sleepiness estimation method according to an embodiment of the present invention includes a cumulative average heart rate obtained by dividing a cumulative total of heart rate acquired in the acquisition step by time after a change in the user's activity state in the cumulative average heart rate calculation step. Based on the accumulated average heart rate in the drowsiness reference heart rate calculation step, a drowsiness reference heart rate that is a reference for whether or not the user feels drowsiness is calculated. In the sleepiness estimation step, the user's current sleepiness is estimated based on the sleepiness reference heart rate and the current heart rate. By doing so, drowsiness can be estimated based on the accumulated heart rate, so the drowsiness reference heart rate can be calculated in a short time, and drowsiness can be estimated with higher accuracy.

  Moreover, it is good also as a sleepiness estimation program which performs the sleepiness estimation method mentioned above by computer. In this way, since sleepiness can be estimated based on the accumulated heart rate using a computer, the sleepiness reference heart rate can be calculated in a short time, and sleepiness can be estimated with higher accuracy.

  A drowsiness estimation apparatus according to a first embodiment of the present invention will be described with reference to FIGS. As shown in FIG. 1, the sleepiness estimation apparatus 1 includes a state determination unit 2, a parameter table 3, a sleepiness reference heart rate calculation unit 4, a sleepiness calculation unit 5, a sleepiness display unit 6, an I / F 7, an I / F F8. The sleepiness estimation apparatus 1 is connected to a heart rate sensor 11.

  The drowsiness estimation apparatus 1 having the configuration shown in FIG. 1 may be configured as an application program (app) such as a mobile device such as a smartphone or a personal computer, or installed in an in-vehicle device such as a car navigation system or a home appliance. May be. In addition, the sleepiness estimation apparatus 1 may include the heart rate sensor 11.

  Note that the sleepiness estimation device 1 including the heart rate sensor 11 as a measurement unit is a state monitoring device 100 that monitors the state of a user (subject), and for example, by mounting the state monitoring device 100 on a vehicle, a driver monitor Can be operated as Of course, the present invention is not limited to the driver monitor, and other activity states such as desk work, study, or work in a factory can be monitored.

  Any known sensor can be used as the heart rate sensor 11 as long as it can acquire at least a heart rate as biological information related to the heart rate. For example, various forms such as a wristwatch type, a type embedded in a vehicle seat, a type provided in an indoor chair or bed, and the like can be used. Further, the heart rate sensor 11 is not limited to one type, and a plurality of types may be used depending on an activity state described later.

  In the sleepiness estimation apparatus 1 having the above-described configuration, the state determination unit 2, the sleepiness calculation unit 5, and the sleepiness reference heart rate calculation unit 4 function as an arithmetic device such as a CPU (Central Processing Unit). The parameter table 3 functions as a storage medium such as a hard disk or a flash memory (the storage medium may be volatile so that it is loaded each time, but is preferably non-volatile). The sleepiness display unit 6 functions as a display device such as a liquid crystal display. The I / Fs 7 and 8 function as a communication control unit that communicates with the heart rate sensor 11 and the like.

  The state discriminating unit 2 discriminates the user's activity state such as rest, driving, working, and sleep based on the user position information and the user acceleration information input from the outside via the I / F 8. That is, the state determination unit 2 detects the activity state of the user. That is, the state determination unit 2 can detect a change in the activity state of the user.

  As the position information, for example, latitude / longitude information obtained by GPS (Global Positioning System) is input. As the acceleration information, information on an acceleration sensor included in the portable device, the vehicle itself, or the vehicle-mounted device is input. Further, information may be acquired from a device that can incorporate an acceleration sensor such as an activity sensor (activity meter) or the like and can detect a human activity state.

  An example of the determination of the activity state in the state determination unit 2 is shown in FIG. FIG. 2 is an example of a table for determining the activity state. In the table of FIG. 2, the content of the state sta1 indicates a resting state, the content of the state sta2 indicates driving, the content of the state sta3 indicates that the operation is in progress, and the content of the state sta4 indicates the active state of sleep. Of course, the activity state may include items other than the illustrated items such as exercise and heavy labor.

  In the table of FIG. 2, when the position information is living, the acceleration information is small, the speed is small, and the device is a wristwatch, the state is determined to be sta1 (rest). Here, the speed may be calculated from a change in position information. Or if it is a vehicle, you may make it acquire from a speed sensor. The device indicates a type of the heart rate sensor 11 or a place where the heart rate sensor 11 is installed. The device may be acquired from the heart rate sensor 11 or may be separately set by a user or the like.

  Further, when the position information is the vehicle, the acceleration information is medium, the speed is large, and the device is the driver's seat, the state is determined as sta2 (during driving). When the position information is the office, the acceleration information is large, the speed is medium, and the device is a chair, the state is determined as sta3 (working). When the position information is the bed, the acceleration information is small, the speed is small, and the device is the bed, it is determined as the state sta4 (sleep). In particular, the acceleration information and the speed are examples, and specific numerical values can be arbitrarily changed on the setting as appropriate.

  In the table of FIG. 2, the determination is made based on the position information, the acceleration information, the speed information, and the device information. However, the determination may be made using only one item, or only two items or three items. Alternatively, the determination may be made based on the heart rate measured by the heart rate sensor 11. For example, when the change in heart rate is small, it is possible to make a rough determination such as resting or sleeping, and when the heart rate is low but there is fluctuation, driving or working.

  The parameter table 3 includes a table in which sleepiness prediction parameters described later are set for each activity state. Based on the determination result of the state determination unit 2, the sleepiness prediction parameter corresponding to the activity state is output to the sleepiness calculation unit 5. An example of the table is shown in FIG.

  As shown in FIG. 3, two kinds of sleepiness prediction parameters a and b are set. In the example of FIG. 3, in the state sta1 (rest), the sleepiness prediction parameter a is 15 and the sleepiness prediction parameter b is 0.2, and in the state sta2-1 (during driving: manual operation), the sleepiness prediction parameter a is 20, the drowsiness prediction parameter b is 0, and in the case of the state sta2-2 (during driving: automatic operation), the drowsiness prediction parameter a is 20 and the drowsiness prediction parameter b is 0.2. In the state sta3 (working), the sleepiness prediction parameter a is 15 and the sleepiness prediction parameter b is 0.2. In the state sta4 (sleep), the sleepiness prediction parameter a is 10 and the sleepiness prediction parameter b is 0. 2. In the example of the table of FIG. 3, values in the range of 0 to 100 are set for the sleepiness prediction parameters a and b.

  In FIG. 3, the sleepiness prediction parameter divides the case where the state is driving (sta2) into manual driving (sta2-1) and automatic driving (sta2-2). This is because, as will be described later in detail, it is necessary to change the setting value of the sleepiness prediction parameter between manual operation and automatic operation.

  The drowsiness reference heart rate calculation unit 4 calculates a drowsiness reference heart rate using a calculation formula described later. Here, the sleepiness reference heart rate is a heart rate when sleepiness occurs. That is, it is a heart rate at which a person feels sleepy below this heart rate. That is, it is a heart rate used as a reference for the user to feel sleepy. Examples of the method for calculating the drowsiness reference heart rate include a method for calculating based on the average heart rate and a method for calculating based on the minimum heart rate.

  Further, the drowsiness reference heart rate may be separately calculated with the configuration shown in FIG. 4 and transferred to the drowsiness estimation apparatus 1. The drowsiness reference heart rate calculation unit 21 calculates the drowsiness reference heart rate using a calculation formula described later.

  Next, calculation of the sleepiness reference heart rate will be described. It is known that the range of heart rates that a person can take is determined, and the heart rate distribution is close to a normal distribution. It is also known that the heart rate is low when feeling sleepy. Therefore, it can be said that the sleepiness reference heart rate, which is the heart rate when sleepiness occurs, is between the lowest heart rate and the average heart rate.

Therefore, in the present embodiment, the cumulative average heart rate, which is the time averaged value of the cumulative number of heart rates acquired by the I / F 7 from the heart rate sensor 11, and the minimum value of the heart rate acquired by the I / F 7 from the heart rate sensor 11. The average value of (minimum heart rate) and the sleepiness reference heart rate is used. That is, assuming that the cumulative average heart rate is HR_ave and the minimum heart rate is HR_min, the sleepiness reference heart rate HR_ref is calculated by the following equation (1).
HR_ref = (HR_ave + HR_min) / 2 (1)

Alternatively, the cumulative average heart rate HR_ave described above may be used as the drowsiness reference heart rate. That is, you may calculate with the following (2) Formula.
HR_ref = HR_ave (2)

  The cumulative average heart rate is preferably calculated after a predetermined time of about 10 minutes elapses after the activity state changes. Immediately after the change of the activity state, the influence of the previous activity state remains on the heart rate. Therefore, the influence of the previous activity state can be reduced by calculating after the elapse of a predetermined time.

  The cumulative average heart rate is preferably initialized when the activity state changes. For example, when taking a break while driving a vehicle or the like, the accumulated average heart rate is once reset at the time of the break and the accumulation is started. At this time, the minimum heart rate value is also reset.

  That is, the sleepiness reference heart rate calculation unit 4 calculates a cumulative average heart rate by dividing a cumulative total heart rate acquired by the I / F 7 (acquisition unit) by time after the user's activity state has changed. It functions as a number calculation means and a drowsiness reference heart rate calculation means for calculating a drowsiness reference heart rate that is a reference for whether or not the user feels drowsiness based on the cumulative average heart rate. Also, it functions as a minimum heart rate determination unit that determines the minimum heart rate that is the minimum value of the heart rate per predetermined time among the heart rates acquired by the I / F 7 (acquisition unit). Also, by functioning as these means, the sleepiness reference heart rate calculation unit 4 executes the cumulative average heart rate calculation step and the sleepiness reference heart rate calculation step.

  Here, the present inventors conducted an experiment to investigate the relationship between the user's average heart rate and minimum heart rate and actual sleepiness. In the experiment, automatic type passenger cars were being driven and passenger seated as active states, and 11 healthy subjects (men aged 30 to 59) were targeted. The subject wore a heart rate meter (MyBeat manufactured by Union Tool Co., Ltd.) and recorded the high frequency fluctuation component (HF) corresponding to the heart rate and respiratory fluctuation. For evaluation of sleepiness intensity, VAS (Visual Analog Scale) was used to record the subjective value of sleepiness intensity felt by the subject.

  The following conditions were determined during the experiment. In order to prevent factors other than driving conditions from affecting the heart rate and sleepiness, we prohibited the use of conversations while riding, listening to music, and guidance by car navigation. Similarly, in order to eliminate the influence of meals, the experiment was conducted between 10:00 am and 12:00 pm, avoiding about 2 hours after breakfast. The subject sitting in the passenger seat was given a simulated role as a supervisor of automatic driving. In order to avoid the tension and stress load, he refrained from changing lanes as much as possible and instructed the left lane to drive in a natural flow.

  As the experimental procedure, the subjects sat in the driver's / passenger seats in pairs, and the driver was changed when the turnaround point arrived. First, to prevent starting the experiment from a sleepy state (to reset the wakefulness and start the experiment), after giving the subject a simple mental arithmetic task for 5 minutes and applying stress, another 5 Rested for a minute. Thereafter, driving on a general road and a course running at a high speed, which will be described later, was started, and during the running, my sleepiness was recorded using VAS at a timing of every 5 minutes when a timer sounded. At the time of the experiment, we ran about 10 km on the general road (required time about 30 minutes), about 40 km on the expressway (required time about 30 minutes), and about 2 hours on a round trip.

5 to 7 show a comparison between the VAS results of three subjects and the sleepiness Sp estimated from the sleepiness reference heart rate calculated by the above-described equations (1) and (2). The upper part of FIGS. 5 to 7 is the result of VAS, and the lower part is sleepiness Sp estimated based on the sleepiness reference heart rate calculated by the above-described equations (1) and (2). In the lower graph, the solid line indicates the case where the drowsiness reference heart rate is calculated by the equation (1), and the alternate long and short dash line indicates the case where the drowsiness reference heart rate is calculated by the equation (2). The sleepiness Sp is expressed by the following equation (3), assuming that the sleepiness prediction parameters are a and b, the heart rate HR currently measured per minute, the sleepiness reference heart rate HR_ref, the high frequency fluctuation component HF, and the reference high frequency fluctuation component HF_ref. Calculated.
Sp = a × max (HR_ref−HR, 0) + b × max (HF−HF_ref, 0) (3)

  The function max is a function whose return value is the maximum value among arguments separated by commas in parentheses. In the expression (3), when HR <HR_ref, the calculation result of HR_ref−HR is HF> In the case of HF_ref, the calculation result of HF-HF_ref becomes a return value.

  In addition, the reference high frequency fluctuation component HF_ref is set in the parameter table 3 for each time or activity state. Details of the high frequency fluctuation component HF and the reference high frequency fluctuation component HF_ref will be described later.

  As shown in FIGS. 5 to 7, it became clear that the sleepiness estimated based on the above-described equations (1) and (2) can estimate sleepiness in the same pattern as the VAS result.

  Moreover, when sleepiness is estimated based on Formula (2), as shown by A in the lower part of FIGS. This occurs before drowsiness becomes larger compared to the VAS in the upper stage and the result, and is considered a sign of drowsiness. Therefore, in the case of the formula (2), it has been clarified that it can be estimated from a sign before the drowsiness is noticed.

  Therefore, the equation (2) is preferably applied in the case of an active state in which the relationship between sleepiness and danger such as driving a vehicle is high. Alternatively, the user or the like may arbitrarily select whether to consider the sign, and the formula (1) or (2) may be switched. The method for calculating the drowsiness reference heart rate described above can be applied not only during driving but also in other activity states.

  In this embodiment, as described above, the sleepiness reference heart rate is calculated by obtaining the cumulative average heart rate and the minimum heart rate in real time from the acquired heart rate and the change in the activity state, so the sleepiness reference heart rate is preset. There is no need. Therefore, for example, by mounting the sleepiness estimation device 1 in a rental car or the like, it is possible to accurately estimate sleepiness for an unspecified number of users according to each user.

  Note that the drowsiness reference heart rate obtained by the above-described equations (1) and (2) may be set as a history in a table or the like for each activity state and used as an initial value or the like.

  The sleepiness calculator 5 is based on the heart rate measured by the heart rate sensor 11, the sleepiness prediction parameter output from the parameter table 3, and the sleepiness reference heart rate output from the sleepiness reference heart rate calculator 4. The user's sleepiness is estimated by calculating (information on sleepiness).

  The sleepiness calculated by the sleepiness calculation unit 5 is information obtained from the difference between the current heart rate and the sleepiness reference heart rate when the current heart rate is lower than the sleepiness reference heart rate. The drowsiness value is calculated so as to increase in accordance with the rate of decrease from the drowsiness reference heart rate.

  For example, people often have strong sleepiness in the afternoon. It is also called post lunch dip because it happens after lunch, and many people think that it is caused by having lunch. However, even if the lunch is advanced by 2 hours or if the lunch is skipped, even if a snack is taken every 2 hours using the constant method, drowsiness occurs in the afternoon. In other words, even if the sleep is not insufficient or the influence of lunch is removed, sleepiness occurs in the afternoon, and it is considered that the afternoon sleepiness reflects biological rhythm.

  Sleepiness is thought to involve a circadian rhythm with a period of about 24 hours, a circasemidian rhythm with a period of about 12 hours, and an ultradian rhythm with a period of about 2 hours. Yes. The nighttime sleepiness is consistent with a decrease in body temperature rhythm and reflects the circadian rhythm of body temperature. Afternoon sleepiness occurs about half a day after the nighttime minimum body temperature appears, and is considered to reflect circadian rhythm.

  Here, the premise for calculating sleepiness will be described. In order to estimate the balance of the autonomic nerve from the heart rate variability, from the time-series data about the heart rate variability, the high frequency variability component (HF component) corresponding to the respiratory variability and the low frequency corresponding to the Mayer wave which is the blood pressure variability A component (LF component) is extracted, and the size of both is compared. The HF component reflecting the respiratory fluctuation appears in the heartbeat fluctuation only when the parasympathetic nerve is in tension (activated). On the other hand, the LF component also appears in heart rate variability when the sympathetic nerve is tense and when the parasympathetic nerve is tense.

The drowsiness calculator 5 calculates the drowsiness S by the following equation (4).
S = Ki × Si + Kp × Sp + Kd × Sd (4)

  In the equation (4), Sp is physiological sleepiness, Si is an integral element of physiological sleepiness Sp, and Sd is a differential element of physiological sleepiness Sp. Ki, Kp, and Kd are coefficients for weighting each element.

  The physiological sleepiness Sp is calculated by the above-described equation (3). The sleepiness calculation unit 5 that calculates the physiological sleepiness Sp calculates a sleepiness value indicating the user's sleepiness based on at least the difference between the user's heart rate and the sleepiness reference heart rate.

  Here, setting the sleepiness prediction parameters a and b will be described. For example, drowsiness during driving tends to occur during monotonous driving (such as an expressway), resulting in functional deterioration accompanied by drowsiness. Such functional degradation appears while the heart rate, blood pressure, etc. are calm in terms of physiological functions, and eye movements and abnormalities of the electroencephalogram are shaken. Subjective symptoms / feeling of fatigue are greatly increased, mainly due to drowsiness and tiredness of the extremities, and a strong decline in concentration is also felt. There is a large extension and variation in behavioral ability and reaction time, and there is a decrease in accuracy, leading to a dangerous situation due to closed eyes and stagnation.

The state of drowsiness is classified as follows according to the function of the autonomic nerve.
<1. If you are not drowsy>
The sympathetic nerve is enhanced and the parasympathetic nerve is suppressed. Heart rate HR is large, and high-frequency fluctuation component (also referred to as high-frequency fluctuation component of heartbeat fluctuation) HF corresponding to respiratory fluctuation is small.
<2. If you have signs of sleepiness>
Due to monotonous driving and fatigue, there is little awareness of sleepiness psychologically, but the signs (predictors) appear physiologically. Since the sympathetic nerve activity changes from the enhanced state to the inhibited state, the heart rate decreases.
<3. When sleepiness occurs>
Although the sympathetic nerve remains suppressed, the parasympathetic nerve activity changes to an enhanced state, so that the heart rate HR decreases and the high-frequency fluctuation component HF corresponding to the respiratory fluctuation increases.
<4. Conflict against sleepiness>
To feel danger and resist sleepiness, create tension. In the case of an accident, the sympathetic nerve activity is intermittently increased, and the high frequency fluctuation component HF corresponding to the respiratory fluctuation is reduced.
<5. State that can not resist sleepiness>
Tension disappears and doze begins. Since the sympathetic nerve activity is suppressed, the heart rate HR is lowered.

  According to the above classification, in normal doze driving, it changes from state 1 (when there is no sleepiness) to state 2 (when there is a sign of sleepiness) and further to state 4 (conflict state against sleepiness). There are many. On the other hand, in the resting state, the state changes from 1 to 2 and becomes 3 (when sleepiness occurs), and when it falls asleep, it reaches 5 (state that cannot resist sleepiness).

  Therefore, in equation (3), as shown in FIG. 3, during manual operation, the sleepiness prediction parameter b of heartbeat fluctuation is set to 0 for calculation. That is, in the manual driving state, the sleepiness prediction parameter a for heart rate change is increased, and the sleepiness prediction parameter b for heart rate fluctuation is decreased. On the other hand, during automatic driving, even if the driver is not driving, since it is close to a resting state, the sleepiness prediction parameter a as an element of the heart rate change is reduced, and sleepiness prediction of heart rate fluctuation is performed. Increase the parameter b.

  That is, the sleepiness prediction parameters a and b are coefficients for weighting the heart rate (RR interval) and heart rate fluctuation. As described above, since the heart rate and heart rate fluctuation have different contributions to sleepiness depending on the activity state, the accuracy of the calculated sleepiness is increased by weighting each.

The integral element Si of the physiological sleepiness Sp is calculated by the following equation (5).
Si = ΣSp (5)

  Equation (5) shows an integral element when sleepiness is accumulated. That is, the integrated value of the sleepiness value (physiological sleepiness Sp) is calculated. For example, in the driving state, drowsiness is low after the start of driving, but drowsiness increases when driving continues for a long time. Therefore, by integrating the physiological sleepiness Sp, sleepiness that occurs when a specific activity state is continued can be taken into consideration. For example, in the driving state, the integration element Si is set to 0 at the start of driving, and then accumulates until the driving is stopped due to arrival at the destination or a break. That is, when the activity state changes, the integrated value is reset.

  The change in the activity state can be detected by the state determination unit 2 as described above. For example, if the drowsiness estimation device is in-vehicle, the change in the activity state may be another activity state when it is detected that the driver has left the vehicle, or when driving is stopped when the vehicle ignition switch is turned off. What is necessary is just to determine with having changed to. Moreover, what is necessary is just to determine with having changed into the driving | running state because the driver | operator etc. sat down in the driver's seat or having turned ON the ignition switch.

  As described above, the integration element Si is integrated indefinitely as long as the specific active state is continued. In this case, depending on the user, there may be a difference between sleepiness S and his / her own sense as time passes. Accordingly, an upper limit value may be provided for the integral element Si so that the value does not exceed the upper limit value. That is, when the upper limit value is exceeded, the upper limit value may be used as the integrated value.

The differential element Sd of the physiological sleepiness Sp is calculated by the following equation (6).
Sd = dSp / dt (6)

  The differential element Sd is calculated to emphasize the change in sleepiness. That is, the differential value of the sleepiness value is calculated. When there is little change in sleepiness, the increase or decrease in sleepiness in a minute interval is easily perceived psychologically (Weber's law). For example, when sleepiness is displayed as a numerical value or the like during driving, the user can be more strongly aware of the change in sleepiness by enlarging a minute change and feeding it back.

  The coefficients Ki, Kp, and Kd can be set for each individual user. The coefficient Kd of the differential element Sd is Kd> 0. Sensitive users increase Kd to make the change easy to feel, and insensitive users make Kd small to make the change difficult to feel. The coefficients Ki, Kp, and Kd may be changed according to the physical condition of the user acquired by the vital sensor or the user's own notification.

  As described above, the differential element Sd has an effect when notifying the user or the like of the display, and is not necessarily an element necessary for calculating the sleepiness S. Therefore, the drowsiness S may be calculated using only the physiological drowsiness Sp and the coefficient Kp, the integral element Si, and the coefficient Ki.

  The sleepiness S calculated by the equation (4) is set to a numerical value range of 0 to 100 so as to be easily compared with the evaluation axis of subjective sleepiness such as VAS. As for this drowsiness S, 0 is small drowsiness and 100 is large drowsiness.

  That is, since the sleepiness prediction parameter varies depending on the activity state, the sleepiness calculation unit 5 calculates the sleepiness S based on the heart rate (biological information), the activity state, and the sleepiness reference heart rate.

  The sleepiness display unit 6 displays the sleepiness S calculated (estimated) by the sleepiness calculation unit 5. The drowsiness S may be simply displayed as a numerical value at that time, or may be displayed as a bar graph or a line graph so that a change in time series can be understood.

  A display example of the sleepiness display unit 6 is shown in FIG. FIG. 8A is a table showing the heart rate at each time, the sleepiness reference heart rate, and the calculated sleepiness S. FIG. 8 (b) is a bar graph of FIG. 8 (a). That is, when the heart rate is measured as shown in FIG. 8A, it is displayed to the user as shown in FIG. 8B. In this display, the drowsiness S including the above-described differential element Sd is displayed, and the drowsiness display unit 6 displays a display indicating drowsiness based on the drowsiness estimated by the drowsiness calculation unit 5 (estimating means). Functions as display control means.

  I / F 7 is an interface (I / F) to which the heart rate sensor 11 is connected. The I / F 7 is an interface corresponding to wired connection when the heart rate sensor 11 is wired connection, and an interface corresponding to wireless connection when the heart rate sensor 11 is wireless connection. That is, the I / F 7 acquires biological information (heart rate, measurement time, etc.) related to the user's heart rate measured by the user. That is, the I / F 7 functions as an acquisition unit that acquires the heart rate and the measurement time.

  The I / F 8 is an interface (I / F) through which position information and acceleration information are input. The I / F 8 may be either a wired connection interface or a wireless connection interface. That is, the I / F 8 acquires user position information and user acceleration information.

  FIG. 9 shows a flowchart of the operation of the sleepiness estimation apparatus 1. FIG. 9A is a flowchart of the overall operation. First, in step S <b> 11, the heart rate is acquired from the heart rate sensor 11, and the state determination unit 2 determines the user's activity state.

  Next, in step S12, based on the heart rate acquired from the heart rate sensor 11, the sleepiness prediction parameter output from the parameter table 3, and the sleepiness reference heart rate output from the sleepiness reference heart rate calculation unit 4, the sleepiness calculation unit 5 To calculate sleepiness S. In step S13, sleepiness S is displayed.

  FIG. 9B is a flowchart of the sleepiness calculation operation in step S12. First, in step S21, physiological drowsiness Sp is calculated by equation (3). Next, in step S22, an integral element Si of the physiological sleepiness Sp is calculated. As described above, the integration element Si is integrated from the start of a specific activity such as driving until it is stopped. Next, a differential element Sd of the physiological sleepiness Sp is calculated. Strictly speaking, the differential element Sd is the expression (6), but in order to simplify the calculation, the Sp calculated based on the previously measured HR or the like from the Sp calculated based on the HR or the like measured this time is used. Is calculated by subtracting. That is, it calculates based on measurement intervals, such as HR. In step S24, drowsiness S is calculated by equation (4). That is, FIG. 9B is a sleepiness estimation step.

  According to the present embodiment, the drowsiness estimation device 1 uses the accumulated average heart rate obtained by dividing the accumulated heart rate acquired by the I / F 7 by the time after the drowsiness reference heart rate calculation unit 4 changes the user's activity state. Based on the accumulated average heart rate, a drowsiness reference heart rate that is a criterion for whether or not the user feels drowsiness is calculated. Then, the sleepiness calculation unit 5 estimates the user's current sleepiness based on the sleepiness reference heart rate and the current heart rate. By doing so, drowsiness can be estimated based on the accumulated heart rate, so the drowsiness reference heart rate can be calculated in a short time, and drowsiness can be estimated with higher accuracy.

  The drowsiness reference heart rate calculation unit 4 calculates the cumulative average heart rate after a predetermined time elapses after the user's activity changes. By doing so, the influence of other activity states before the change can be reduced, and the calculation accuracy of the sleepiness reference heart rate can be improved.

  Further, the sleepiness reference heart rate calculation unit 4 determines the lowest heart rate that is the lowest value of the heart rate per predetermined time acquired by the I / F 7, and calculates the average value of the cumulative average heart rate and the minimum heart rate as sleepiness. Calculated as the reference heart rate. By doing so, the mechanism for calculating the drowsiness reference heart rate is simplified, so that it is possible to reduce processing time and development costs in the apparatus and the like.

  The drowsiness reference heart rate calculation unit 4 calculates the cumulative average heart rate as the drowsiness reference heart rate. By doing so, the mechanism for calculating the drowsiness reference heart rate can be further simplified, and the processing time and development cost in the drowsiness estimation device 1 can be reduced. In addition, by setting the cumulative average heart rate as the drowsiness reference heart rate, it is possible to capture a sign of feeling drowsiness, and to take a measure such as prompting a break before recognizing drowsiness.

  In addition, the sleepiness reference heart rate calculation unit 4 initializes the accumulated heart rate when the activity state changes. By doing so, it is possible to avoid being affected by changes in sleepiness caused by a break or when the day changes.

  In addition, the drowsiness reference heart rate calculation unit 4 may calculate the cumulative average heart rate after the user changes to the activity of driving the mobile object. In this way, when driving a moving body such as a vehicle, the drowsiness reference heart rate can be calculated in a short time, and drowsiness can be estimated with high accuracy.

  The sleepiness calculator 5 calculates the sleepiness S of the user based on the sleepiness prediction parameter set in advance for each activity state. In this way, the heart rate and heart rate fluctuation can be weighted according to the activity state.

  Moreover, I / F8 which acquires a user's positional information and acceleration information is provided. In this way, the activity state can be detected from the moving speed and acceleration of the user.

  Further, since the drowsiness display unit 6 is provided, the user can specifically perceive his / her drowsiness, and can take a break or exercise, for example, based on the displayed drowsiness.

  In the above-described embodiment, the sleepiness prediction parameter is selected based on the activity state and the measurement time. However, the sleepiness prediction parameter may be selected based only on the activity state. However, it is preferable to consider the measurement time because an appropriate sleepiness prediction parameter can be selected.

  Further, sleepiness may be notified by voice or vibration in addition to the sleepiness display unit 6. Alternatively, notification by voice or the like may be performed in the case of drowsiness above a certain level.

  If there is a difference between the sleepiness value calculated (estimated) in the sleepiness estimation apparatus 1 having the above-described configuration and the user interval, a subjective evaluation is input, and the sleepiness criterion is based on the subjective evaluation. The heart rate HR_ref and the reference high frequency fluctuation component HF_ref may be corrected. By doing so, the user can subjectively evaluate the sleepiness calculated (estimated) by the sleepiness calculation unit 5. Then, the drowsiness calculation unit 5 can calculate drowsiness again by feeding back the subjective evaluation result. Therefore, it is possible to calculate information regarding sleepiness tailored to the individual user with higher accuracy.

  Next, a display device according to a second embodiment of the present invention will be described with reference to FIGS. The same parts as those in the first embodiment described above are denoted by the same reference numerals and description thereof is omitted.

  In the present embodiment, as shown in FIG. 10, the parameter table 31 is provided in the server 30. The server 30 also stores individual history data 32. The server 30 can collectively manage data of a plurality of persons.

  In the individual history data 32, history data such as heart rate, measurement time, activity state, etc. received from the device 1A described later is stored for each individual. The server 30 transmits a reference high frequency fluctuation component, a drowsiness prediction parameter, and the like based on a request from the device 1A or the like.

  The device 1A and the device 1B are devices that are installed in the vicinity of the user or worn by the user, can acquire at least history data, and can transmit the acquired history data to the server 30. It has become.

  In FIG. 10, for example, a wristwatch type and an in-vehicle type are shown as the device 1 </ b> A and the device 1 </ b> B, but these can be used by the same user and the history data acquired by each can be transmitted to the server 30. Thus, even if the same user uses different devices 1A and 1B, history data can be stored in one place, and the parameter table 31 can be shared by a plurality of devices 1A and devices 1B. .

  FIG. 11 shows a device 1A according to the present example. The device 1B basically has the same configuration. In FIG. 11, the state determination unit 2, the drowsiness display unit 6, and the I / F 7 are basically the same as those in the first embodiment. However, the state determination unit 2 is different in that the determined state is output to the communication unit 12. The sleepiness reference heart rate / sleepiness calculation unit 5A calculates a sleepiness reference heart rate based on the heart rate acquired by the I / F 7, and based on the sleepiness reference heart rate and the sleepiness prediction parameter acquired by the communication unit 12. Calculate sleepiness. The I / F 7 outputs the acquired heart rate and measurement time to the communication unit 12 in addition to the state determination unit 2 and the drowsiness calculation unit 5.

  The communication unit 12 communicates with the server 30. The communication unit 12 transmits the heart rate measured by the heart rate sensor 11, the measurement time, and the activity state determined by the state determination unit 2 to the server 30, and receives the sleepiness prediction parameter transmitted from the server 30.

  In the present embodiment, the device 1A (device 1B) may be provided with the heart rate sensor 11. The server 30 may include the state determination unit 2 and the drowsiness reference heart rate / drowsiness calculation unit 5A. In this case, the communication unit 12 transmits information for determining an activity state such as position information and acceleration information. That is, the server 30 may determine the activity state and calculate the sleepiness S. Alternatively, the server 30 may include only the state determination unit 2 or the drowsiness reference heart rate / drowsiness calculation unit 5A.

  That is, the sleepiness estimation apparatus according to the present embodiment includes the server 30 and the device 1A (device 1B).

  According to the present embodiment, since the sleepiness estimation apparatus is configured by the server and the device 1A (device 1B), even when one user uses a plurality of devices, history data such as heart rate and sleepiness reference heartbeats are used. Numbers can be managed collectively. Further, the parameter table can be shared by a plurality of devices 1A (device 1B).

  In addition, this invention is not limited to the said Example. That is, those skilled in the art can implement various modifications in accordance with conventionally known knowledge without departing from the scope of the present invention. Of course, such modifications are included in the scope of the present invention as long as the configuration of the sleepiness estimation apparatus of the present invention is provided.

1, 1A Drowsiness estimation device 5 Drowsiness calculation unit (sleepiness estimation means)
7 I / F (Acquisition means)
9 Drowsiness reference heart rate calculation unit (cumulative average heart rate calculation means, drowsiness reference heart rate calculation means, minimum heart rate determination means)
11 Heart rate sensor (measuring means)
100 Condition monitoring device

Claims (8)

Acquisition means for acquiring a user's heart rate;
A cumulative average heart rate calculation means for calculating a cumulative average heart rate obtained by dividing a cumulative total of heart rates acquired by the acquisition means after time since the user's activity state changed;
Drowsiness reference heart rate calculation means for calculating a drowsiness reference heart rate that is a reference as to whether or not the user feels drowsiness based on the cumulative average heart rate;
Sleepiness estimation means for estimating the current sleepiness of the user based on the sleepiness reference heart rate and the current heart rate;
A drowsiness estimation device comprising:   The sleepiness estimation apparatus according to claim 1, wherein the cumulative average heart rate calculation means calculates the cumulative average heart rate after a predetermined time has elapsed since the activity state has changed. Further comprising a minimum heart rate determination means for determining a minimum heart rate that is a minimum value of the heart rate acquired by the acquisition means;
3. The sleepiness estimation apparatus according to claim 1, wherein the sleepiness reference heart rate calculation unit calculates an average value of the cumulative average heart rate and the lowest heart rate as the sleepiness reference heart rate.   The sleepiness estimation apparatus according to claim 1 or 2, wherein the sleepiness reference heart rate calculation means calculates the cumulative average heart rate as the sleepiness reference heart rate.   5. The cumulative average heart rate calculation unit calculates the cumulative average heart rate after the user has changed to an active state of driving a moving body. 6. Sleepiness estimation device. Drowsiness estimation device according to any one of claims 1 to 5,
Measuring means for measuring the user's heart rate;
A state monitoring device comprising: A sleepiness estimation method for a sleepiness estimation device that estimates a user's current sleepiness,
An acquisition step of acquiring a user's heart rate;
A cumulative average heart rate calculating step of calculating a cumulative average heart rate obtained by dividing a cumulative total of heart rates acquired by the acquisition step by time since the user's activity state changed;
Based on the cumulative average heart rate, a sleepiness reference heart rate calculation step of calculating a sleepiness reference heart rate that is a reference for whether or not the user feels sleepiness;
A sleepiness estimation step of estimating the current sleepiness of the user based on the sleepiness reference heart rate and the current heart rate;
A drowsiness estimation method comprising:   A sleepiness estimation program characterized by causing a computer to execute the sleepiness estimation method according to claim 7.

Images