JP2007044154A - Awaking-degree judging method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an awaking-degree judging method for judging an awaking degree based on the fluctuation of the intervals of heart beat by which the awaking-degree can be precisely measured even when the abrupt mental load fluctuation is generated. <P>SOLUTION: The number of samples is determined so that heart beat frequency band in which the appearance of breath fluctuation of the intrapulse time (RRI) of the heart beat is expected is present in a passing through zone of a dispersion ratio-the heart beat frequency property, and the RRI is obtained in the frequency band in which the appearance of the breath fluctuation is expected in the number of samples. Next, that the fluctuation condition of the RRI is a normal driving condition, simple increase, or a simple decrease is detected, and a datum line f(n) is changed to one corresponding to the fluctuation of the RRI when the fluctuation of the RRI is the normal driving condition, and when the same is the simple increase or the simple decrease. Thereafter, the dispersion RRVn RRI relative to the datum line f(n) is calculated, and the awaking-degree is detected based on the dispersion RRVn. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an arousal level determination method, and more particularly to an arousal level determination method for determining an arousal level based on variations in heartbeat intervals.

  In order to avoid the danger of driving the vehicle when the driver of the vehicle unexpectedly falls asleep, a method for detecting a decrease in the driver's arousal level has been developed. For example, Patent Document 1 is known as an apparatus / method that has been proposed so far as such a degree of arousal determination.

  The principle of the arousal level determination method disclosed in Patent Document 1 will be described.

  In order to perform arousal level determination using the method disclosed in Patent Document 1, when it is expected that respiratory variation of RRI appears in the heart rate frequency band near the frequency f1, this band is the dispersion ratio-heart rate frequency. The number of samples related to the calculation of RRV is set so as to be included in the characteristic passband. Here, RRI (R-R Interval) is the time between heartbeats, that is, the interval time of the R wave of the electrocardiogram, and RRV indicates the dispersion of RRI at awakening.

  Now, if it is expected that respiratory fluctuations appear at the frequency f1 shown in FIG. 1A, first, as the number of samples n, as shown in FIG. 1B, the band to which this frequency f1 belongs. Is set to the number of samples n1 corresponding to the dispersion ratio (pass ratio relative to the input of RRV) -heart rate frequency characteristics included in the passband.

FIG. 1B shows a dispersion ratio-heart rate frequency characteristic using the number of samples n as a parameter.
n1 <n2 <n3 (1)
The dispersion ratio-heart rate frequency characteristics are shown for three types of samples having the relationship: As shown in this figure, the cut-off frequency increases as the number of samples n decreases. The reason why the dispersion ratio-heart rate frequency characteristic has a high-pass characteristic and the cut-off frequency depends on the number of samples is as follows.

  FIG. 2 shows the relationship between the number of samples and RRV. Since RRV is a variance of RRI as described above, this can be expressed by a mathematical formula:

ここで、RRIiはRRIの測定値、AveRRIは１〜ｎ個のRRIの平均値である。そして、この平均値AveRRIの値を基準線とし、この基準線からの偏差であるAveRRI−RRIiを各サンプルにおいて行うことにより、分散値RRVを求める。

Here, RRIi is a measured value of RRI, and AveRRI is an average value of 1 to n RRIs. Then, the average value AveRRI is used as a reference line, and AveRRI-RRIi, which is a deviation from the reference line, is performed on each sample to obtain a variance value RRV.

  As shown in FIG. 2A, when the number n of samples related to the RRV calculation is sufficiently large, the average value AveRRI is a value reflecting the influence of the low frequency component of the RRI fluctuation. Conversely, when the number of samples n is small, as shown in FIG. 2B, the average value AveRRI does not reflect the low-frequency component, and is the deviation of this average value AveRRI (AveRRI-RRIi). This means that the low frequency component of the RRI fluctuation is not affected.

  For this reason, in the RRV calculation based on Expression (2), the same result as that calculated for the high-pass filter waveform as shown in FIG. 2C is expected. Thus, since the average value AveRRI depends on the number of samples n, the high-pass characteristic of the dispersion ratio-frequency characteristic depends on the number of samples n. That is, as the number of samples n is larger, the influence of the low frequency component of RRI variation is reflected in the RRV calculation.

  Therefore, the number of samples is set to, for example, n1 so that the dispersion ratio-frequency characteristic includes the frequency f1 of the respiratory variation of RRI, and the dispersion ratio-frequency characteristic related to the number of samples n1 is measured, for example, during RRI It can be seen that by multiplying the fluctuation intensity-heart rate characteristic, the respiratory fluctuation at the frequency f1 can be detected as shown in FIG. This can actually be achieved by calculating RRV with the number of samples n1.

  Based on such a principle, first, the number of samples n is set as described above. Next, AveRRI, which is a determination value of RRV corresponding to the permissible limit of wakefulness, is set based on RRV at the time of wakeup based on the number of samples n. Further, RRV is calculated for the number of samples n. Then, the determination value AveRRI and the current RRV are compared (see the above equation (2)), and it is determined whether or not the current arousal level is below the allowable limit.

Accordingly, the number of samples n is set so that the frequency band in which respiratory fluctuations are expected is included in the pass band of the dispersion ratio-frequency characteristic. Therefore, the number of samples n is significantly smaller than the number of data required for FFT analysis, for example. From the principle of 2, n is about the period of respiratory change (about 3 to 5). For this reason, when the arousal level is instantaneously decreased or when the arousal level is monotonously decreased, the respiratory variation is followed and extracted in real time.
Japanese Patent Laid-Open No. 03-272745

  As described above, according to the invention disclosed in Patent Document 1, the arousal level can be determined with a small number of samples. By the way, when the present inventor measured the time between heartbeats in various mental states of the driver of the vehicle, when a sudden change in mental load occurs, the change in heartbeat interval (RRI) is simply reduced or reduced. It turned out that it might increase. This will be described with reference to FIG. 3 and FIG.

  FIG. 3 shows a normal operation state. As described above, in the normal operation state, the low frequency component and the high frequency component are superimposed as shown in FIG. The RRI variance RRV at this time can be accurately obtained by using the average value AveRRI of the RRI of n samples as the determination value.

  FIG. 5 is an enlarged view of the area indicated by the arrow A1 in FIG. In the case shown in the figure, the number of samples is 4, and the AveRRI is AveRRI = (RRI1 + RRI2 + RRI3 + RRI4) / 4. Also, in normal operation, high-frequency components are superimposed on the RRI characteristics, so each value of RRI1, RRI2, RRI3, and RRI4 is a characteristic in which larger and smaller values are repeated alternately around AveRRI. Become. Therefore, the values of variance (hereinafter referred to as variance values) RRV1, RRV2, RRV3, and RRV4 are values that accurately reflect the variance of RRI.

  On the other hand, FIG. 4 shows changes in RRI when a sudden change in mental load occurs. As shown in the figure, when a sudden change in mental load occurs, the driver becomes high tension itself and the RRI is shortened. On the other hand, as shown in FIG. 4, it was found that when a sudden change in mental load occurs, there are many drivers who exhibit RRI characteristics in which the high-frequency component decreases and the low-frequency component mainly. Therefore, when such a sudden change in mental load occurs, there is a region where the RRI characteristic simply decreases as shown by an arrow a in FIG. 4, and a region where the RRI indicated by an arrow b in FIG. 4 simply increases. It becomes the characteristic to do.

  In this way, when RRI simply decreases or increases, it is judged that there is a large variance in the method of calculating the variance value RRV based on AveRRI as in the past, but in reality RRI simply decreases or simply increases. The variance of RRI excluding this simple decrease or simple increase is small. Therefore, if the conventional arousal level determination method is used when the RRI simply decreases or increases, it is impossible to obtain an accurate distribution of the RRI. This will be described with reference to FIG.

  FIG. 6 is an enlarged view of the area indicated by the arrow A2 in FIG. 4 (a part of the area a that simply decreases). Also in the case shown in the figure, the number of samples is 4, so AveRRI is AveRRI = (RRI1 + RRI2 + RRI3 + RRI4) / 4. On the other hand, when a sudden change in mental load occurs, the high-frequency component hardly superimposes on the RRI characteristics, so that each value of RRI1, RRI2, RRI3, RRI4 is a linear function equation f ( n) The characteristic is scattered around the center.

  Therefore, in the method for obtaining the variance for AveRRI as in the conventional arousal level determination method and performing the arousal level determination based on this, although it is actually aggregated into the linear function formula f (n) and the variance is small, A large variance will be shown for AveRRI. For this reason, the conventional wakefulness level determination method cannot perform highly accurate wakefulness level determination, and there is a risk of erroneously determining that it is a nap, although it is not a nap.

  The present invention has been made in view of the above points, and an object thereof is to provide a wakefulness determination method capable of accurately determining wakefulness even when a sudden change in mental load occurs. .

In order to solve the above problem, in the arousal level determination method according to the present invention,
The number of samples is determined so that the heart rate frequency band where the occurrence of respiratory change in the time between heartbeats is expected to be included in the passband of the dispersion ratio-heart rate frequency characteristic, and the appearance of the respiratory change within the number of samples is expected. Obtaining the time between beats of the heartbeat in the frequency band to be
It is detected whether the fluctuation state of the beat time of the heartbeat is a normal driving state, a simple increase or a simple decrease, and when the fluctuation of the beat time of the heartbeat is a normal driving state, the beat of the heartbeat When the variation in inter-time is a simple increase or a simple decrease, the reference line indicating the variation in the time between beats of the heartbeat is changed to one corresponding to the variation,
A variance of the time between beats of the heartbeat with respect to the reference line is calculated, and a wakefulness level is detected based on the variance.

  According to the above invention, when there is a sudden change in mental load on the driver, and the heartbeat interval simply increases or decreases, the judgment value changes corresponding to the simple increase or simple decrease. Therefore, it is possible to determine the arousal level with high accuracy.

  In the above invention, the reference line is a line segment connecting a beat time corresponding to the first sample of the predetermined number of samples and a beat time corresponding to the last sample of the number of samples. Can do.

  In the above invention, the reference line may be a line segment obtained by a least square method based on a time value between beats corresponding to each sample of the determined number of samples.

  According to the present invention, when the time between heartbeats at the time of awakening is simply increased or simply decreased, the determination value is set based on a reference line that changes in response to the simple increase or the simple decrease, so that a highly accurate awakening Degree determination can be performed.

  Next, the best mode for carrying out the present invention will be described with reference to the drawings.

  FIG. 7 shows the configuration of a wakefulness determination device that employs a wakefulness determination method according to an embodiment of the present invention. As shown in the figure, the arousal level determination device has an oscillation circuit 10 that oscillates at a predetermined frequency, and an ultrasonic sensor 14 that transmits an ultrasonic wave by the output of the oscillation circuit 10 and receives a reflected wave. ing.

  The ultrasonic sensor 14 is built in the seat 16 on which the driver 12 sits, and detects the pulsation of the heart of the driver 12, that is, the heartbeat by reflection of ultrasonic waves from the heart. The ultrasonic sensor 14 is connected to an amplifier 18, and the amplifier 18 is connected to a comparator 22 through a filter 20. That is, the heartbeat detected as a reflected wave by the ultrasonic sensor 14 is amplified by the amplifier 18, noise is removed by the filter 20, and converted into a pulse signal by the comparator 22.

  The comparator 22 is connected to the microcomputer 24, and the microcomputer 24 is connected to the awakening means 26. The microcomputer 24 is a device that performs wakefulness determination based on the RRI variance RRV that is the time between heartbeats. The microcomputer 24 takes in information related to the heartbeat by using a pulse signal from the comparator 22 and wakes up means 26 according to the determination result. To work. The awakening means 26 urges the driver 12 to awaken, for example, issues a warning to the driver 12 in response to a command from the microcomputer 24, and improves the arousal level.

  FIG. 8 is a flowchart showing the operation of the arousal determination process performed by the ichrocomputer 24. When this awakening determination process is started, first, in step 10 (step is abbreviated as S in the figure), RRI that is the time between heartbeats (interval time of R wave in the electrogram) is measured. This RRI is information relating to a heartbeat of a predetermined number of beats at awakening, and is detected from the ultrasonic sensor 14 immediately before the vehicle starts running, for example. The predetermined number of beats, that is, the number of samples n, is such that the frequency band in which the RRI respiratory change of the driver 12 is expected, for example, 0.2 to 0.3 [1 / beat] belongs to the dispersion ratio-frequency characteristic pass band. It is determined. In this embodiment, 4 is adopted as the number of samples n.

  When RRI is measured in step 10, it is subsequently determined in step 12 whether the change in RRI is a simple increase or a simple decrease. This change in RRI is judged to be a simple decrease when the values of Sub (n) = RRIn + 1−RRIn are n = 1 to 4 and are all negative (negative) (hereinafter, this state is simply decreased). Called state). Conversely, when the values of Sub (n) are n = 1 to 4 and all are positive (positive), it is determined that the value is simply increased (hereinafter, this state is simply referred to as a simply increased state). FIG. 9 shows an example where the RRI simply decreases, which is equivalent to the region indicated by the arrow A2 in FIG.

  As described above, when the driver has an abrupt mental load, the RRI exhibits a characteristic of simple decrease or increase. On the other hand, when the driver is driving normally (hereinafter, this state is referred to as a normal state), such a simple decrease or increase of RRI does not occur. In addition, when the RRI is in a simple increase state or a simple increase state, using the above-described equation (2) as in the past, the method for obtaining the variance RRV using the AveRRI value, which is the average value of RRI1 to RRI4, as the reference line, As described with reference to FIG. 6, since a simple decrease or a simple increase is not taken into consideration, a large dispersion value is shown, and it is not possible to make a highly accurate arousal level determination.

  In contrast, in this embodiment, it is determined in step 12 whether the change in RRI is a simple increase or a simple decrease. When a negative determination is made in step 12, that is, when it is determined that the current state is normal, the straight line of the AveRRI value is used as a reference line as in the conventional case (see FIG. 5), based on this (2) The variance value RRVn is obtained from the equation (step 16).

  On the other hand, if an affirmative determination is made in step 12, that is, if it is determined that the change in RRI is a simple increase or a simple decrease, the reference line is changed in step 14. FIG. 9 shows an example of the procedure for changing the reference line in step 14. In the figure, the region indicated by the arrow A2 of the simple decrease region a shown in FIG. 4 is shown as an example.

  In step 14, a reference line f (n) that changes corresponding to the simple decrease is obtained. Specifically, in this embodiment, among four samples, the beat time RRI1 corresponding to the first sample (n = 1) and the beat time RRI4 corresponding to the last sample (n = 4) are connected. It is a line segment. Therefore, this reference line f (n) is expressed by the following equation.

f (n) = {(RRI4−RRI1) / 3} × n + a (3)
However, a is a constant.

  The reference line f (n) obtained in this way is a reference line corresponding to a simply decreasing RRI change. Even in the case of a simple increase, a reference line corresponding to the simple increase can be obtained by the same method as described above.

  The method for obtaining the reference line f (n) is not limited to the above method, and other methods can be used. For example, (time (t))-(RRI) coordinates corresponding to four samples may be obtained, and the reference line f (t) may be obtained using the least square method based on the coordinates.

  As described above, when the reference line is changed from the AveRRI that is the reference line during normal operation to the reference line f (n) in step 14, subsequently, in step 16, the variance value RRVn is calculated based on the following equation. The

上記のようにステップ１６で分散値RRVnが演算されると、続くステップ１８ではこの分散値RRVnが居眠り判定値を超えたか否かが判断される。このステップ１８では、先ず居眠り判定値が設定される。この居眠り判定値は、ステップ１６で演算されたRRVに所定の係数αを乗ずることにより行う。この係数αは、予め実験により求められる値であり、ステップ１６で演算されるRRVnがこのRRV×αを超えた場合には、運転者は居眠り状態（覚醒状態）である可能性が高いと判定することができる。

When the variance value RRVn is calculated in step 16 as described above, in the following step 18, it is determined whether or not the variance value RRVn has exceeded the doze determination value. In this step 18, first, a dozing determination value is set. This doze determination value is obtained by multiplying the RRV calculated in step 16 by a predetermined coefficient α. The coefficient α is a value obtained in advance by experiment. When the RRVn calculated in step 16 exceeds this RRV × α, it is determined that the driver is likely to be in a dozing state (wake state). can do.

  Next, in step 18, a comparison process between the dozing determination value (RRV × α) obtained as described above and RRVn obtained in step 16 is performed. In this embodiment, since the sample number n is set so as to pass the respiratory fluctuation as described above, if the driver 12's consciousness is lowered and the arousal level is lowered at this time, the step 18 is performed. RRVn is a large value. Therefore, when this RRVn becomes larger than the dozing determination value, it can be regarded that “the driver 12 is in a dozing state”.

  As described above, the coefficient α defines the determination boundary whether or not “being a doze”. The dormancy determination value obtained by multiplying this coefficient α by the RRVn at the time of awakening is the current RRVn As an indicator of whether or not If the doze determination value is equal to or greater than the current RRVn in step 18, it can be said that the driver's arousal level has not decreased until it can be said that the driver is in a “sleeping state”. Therefore, at this time, the process returns to step 10 again. , Steps 10 to 18 are repeated.

  On the other hand, if the current RRVn exceeds the determination value, it can be regarded as “sleeping state”, so the process proceeds to step 20, and the microcomputer 24 operates the awakening means 26. Thereby, an alarm provided in the driver's seat is activated, and the awakening device is activated. As a result, the driver 12 can be awakened from drowsiness, and safety can be improved. After this, the process again moves to step 10, and the above-described process is repeated.

  As described above, in this embodiment, it is determined in step 12 whether the RRI is simply increased or decreased. If the RRI is not in the normal operation state, RRVn is calculated based on the reference line f (n) obtained by the equation (3). Therefore, it is possible to obtain an accurate RRVn corresponding to the change in the mental load of the driver, and thus it is possible to determine the arousal level with high accuracy.

1A and 1B are diagrams illustrating the principle of extraction of respiratory variation in the arousal level determination method, in which FIG. 1A is an RRI variation intensity-frequency characteristic diagram in which respiratory variation appears, and FIG. 1B is a dispersion ratio indicating a high-pass characteristic. -Frequency characteristic diagram, (C) is an RRI fluctuation intensity-frequency characteristic chart from which respiratory fluctuations are extracted. FIG. 2 is a diagram showing the relationship between the dispersion of the time between heartbeats and the number of samples, (A) is a diagram showing the deviation between the time between heartbeats and the average when the number of samples is relatively large, (B) is a diagram showing the deviation between the average time between heartbeats and the average when the number of samples is relatively small, and (C) is a diagram showing the waveform after passing through the high-pass filter corresponding to FIG. 2 (B). is there. FIG. 3 is a diagram showing the characteristics of the time between heartbeats in the normal driving state. FIG. 4 is a graph showing the characteristics of the time between heartbeats when there is a sudden change in mental load. FIG. 5 is a diagram for explaining how to obtain the dispersion value in the normal operation state. FIG. 6 is a diagram for explaining a problem in the case where the conventional arousal level determination method is used when there is a sudden change in mental load. FIG. 7 is a block diagram illustrating a configuration of an apparatus that employs the arousal level determination method according to an embodiment of the present invention. FIG. 8 is a flowchart for executing the arousal level determination method according to an embodiment of the present invention. FIG. 9 is a diagram for explaining how to obtain a variance value when using the arousal level determination method according to an embodiment of the present invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Transmission circuit 14 Ultrasonic sensor 24 Microcomputer 26 Awakening means

Claims (3)

The number of samples is determined so that the heart rate frequency band where the occurrence of respiratory change in the time between heartbeats is expected to be included in the passband of the dispersion ratio-heart rate frequency characteristic, and the appearance of the respiratory change within the number of samples is expected. Obtaining the time between beats of the heartbeat in the frequency band to be
It is detected whether the fluctuation state of the beat time of the heartbeat is a normal driving state, a simple increase or a simple decrease, and when the fluctuation of the beat time of the heartbeat is a normal driving state, the beat of the heartbeat When the variation in inter-time is a simple increase or a simple decrease, the reference line indicating the variation in the time between beats of the heartbeat is changed to one corresponding to the variation,
A wakefulness level determination method, comprising: calculating a variance of the time between beats of the heartbeat with respect to the reference line, and detecting the wakefulness level based on the variance.   The reference line is a line segment connecting a beat time corresponding to a first sample of the predetermined number of samples and a beat time corresponding to a last sample of the number of samples. The method of determining arousal level according to 1.   2. The arousal level determination method according to claim 1, wherein the reference line is a line segment obtained by a least square method based on a time value between beats corresponding to each of the predetermined number of samples.

Images