JP2943295B2 - Judging method for awakening degree - Google Patents

Description

The present invention relates to a method for determining whether or not a driver's arousal level during driving of a vehicle has fallen below an allowable limit, that is, a method for determining arousal level.

[Prior Art] In order to avoid a danger in running a vehicle when a driver of the vehicle suddenly falls asleep, for example, means for detecting a decrease in arousal level of the driver and issuing an alarm has been developed. .

As an apparatus / method that has been proposed as such means, there is, for example, one disclosed in Japanese Unexamined Patent Application Publication No. 1-1131648 as an “awakening degree determination apparatus”. This apparatus performs an FFT (Fast Fourier Transform) analysis of a fluctuation spectrum of an inter-beat time of a heartbeat, that is, an R-interval (R-R Interval; hereinafter, abbreviated as RRI) of an electrocardiogram, and obtains an RRI of 0.1 [1 / bea.
[t] is a device that extracts fluctuations (so-called blood pressure fluctuations) in the vicinity.

 Here, the fluctuation spectrum of the RRI will be described.

FIGS. 14 to 16 show the fluctuation spectrum of the RRI.

In these figures, the horizontal axis represents the heartbeat frequency [1 / beat] obtained as the reciprocal of the heart rate, and the vertical axis represents the amplitude ratio [% / MAX] representing the fluctuation intensity of the heartbeat interval (RRI). The data of the polygonal line indicates the frequency spectrum of the amplitude ratio.

14 and 15 show peaks not seen in FIG. That is, in FIG. 14, the heart rate frequency is 0.2
A peak is seen near [1 / beat] and in FIG. 15 around 0.25 [1 / beat]. On the other hand, in FIG. 16, such a peak does not appear.

It is known that such a peak appears in deep relation to the arousal state of the subject, and is generally called respiratory fluctuation. The difference between FIG. 14 and FIG. 15 and FIG. 16 is that the former two spectra are when the eyes are closed and the latter is the tension spectrum. In other words, the respiratory variation becomes smaller when nervous, and becomes larger when the eyes are closed.

Further, the frequency of the respiratory variation is represented by respiratory rate / heart rate, and is a value that varies depending on individual differences and the like. The differences between FIG. 14 and FIG. 15 are differences caused by individual differences.

The region where the frequency of the respiratory fluctuation can be taken is a region in the vicinity of 0.2 to 0.3 [1 / beat] based on the result of the experiment by the inventor.

The respiratory variation of the RRI is associated with the driver's arousal level during driving the vehicle. That is, when the degree of arousal is high and the person is nervous, the spectrum becomes as shown in FIG. 16, and the respiratory fluctuation becomes small. Conversely, when the arousal level is low, such as when falling asleep, the spectrum becomes as shown in FIGS. 14 and 15, and large respiratory fluctuations appear.

Therefore, if a large respiratory variation is detected from the frequency spectrum of the RRI, it can be regarded that the subject is in a dozing state or a consciousness decreasing state close to the dozing state.

Further, as is particularly prominently shown in FIG. 15, the RRI spectrum has a peak near 0.1 [1 / beat] due to blood pressure. This peak is called a blood pressure fluctuation, and is an object to be extracted in the above-mentioned Japanese Patent Laid-Open Publication No. 1-1131648.
Blood pressure fluctuations also appear with a decrease in arousal level, similar to respiratory fluctuations. The present inventor has obtained experimental results that show that in the comparison between blood pressure variability and respiratory variability, the latter more sensitively reflects a decrease in arousal level than the former.

Conventionally, the frequency spectrum of the RRI is subjected to FFT analysis to obtain a change from an initial value (at the time of awakening) of the variance in the vicinity of the frequency at which the blood pressure fluctuation is expected, and when this value increases, there is blood pressure fluctuation. Was considered. Then, when it is determined that there is a blood pressure change, an alarm is issued, and the driver of the vehicle is awakened to improve the stability of traveling.

[Problem to be Solved by the Invention] However, performing FFT analysis requires data of a certain number of beats. Usually, the number of beats is on the order of tens.

Therefore, when the time related to the decrease in the arousal level is shorter than the time corresponding to the number of beats required for the FFT analysis, that is, when the arousal level decreases instantaneously, the blood pressure fluctuation or the respiratory fluctuation related to the decrease is recognized. Not detectable.

Also, due to the fact that the FFT analysis is for obtaining an average frequency spectrum during a predetermined number of beats, for example, when the arousal level monotonously decreases, it is useful for detecting blood pressure fluctuation or respiratory fluctuation. A time delay occurs.

These are situations that pose a problem in terms of the safety of driving the vehicle.

Further, as an apparatus for performing the FFT analysis, for example, a microcomputer can be used. However, since such an analysis is complicated, the apparatus and program software become large, complicated, and expensive.

SUMMARY OF THE INVENTION The present invention has been made to solve such a problem, and has an object to improve a follow-up property to a decrease in arousal level, and can be implemented by a simple and inexpensive apparatus. The purpose is to provide.

[Means for Solving the Problems] In order to achieve such an object, the arousal level determination method of the present invention provides a method for determining a respiratory variability of RRI, or a heartbeat expected to exhibit respiratory variability and blood pressure variability. The number of samples is determined so that the frequency band is included in the passband of the variance ratio-heart rate frequency characteristic, and the RR during awakening in the frequency band in which respiratory fluctuation or the appearance of respiratory fluctuation and blood pressure fluctuation is expected
Based on the variance of I (hereinafter abbreviated as RRV), set the judgment value of RRV corresponding to the permissible lowering limit of arousal level, calculate the RRV for the number of samples, and calculate the respiratory fluctuation or respiratory fluctuation of RRI and blood pressure Gender variation is extracted, and the determination value is compared with the current respiratory variation or the respiratory variation and blood pressure variation of the RRI to determine that the current arousal level has fallen below an allowable limit.

In the present invention, the first sample number is determined so that the heartbeat frequency band in which the respiratory variation of the RRI is expected to be included in the passband of the corresponding dispersion ratio-heartbeat frequency characteristic. The second sample number is determined so that the heart rate frequency band in which respiratory fluctuation and blood pressure fluctuation are expected to be included in the pass band of the corresponding variance ratio-heart rate frequency characteristic, and the appearance of respiratory fluctuation is expected. Awake in the frequency band
Based on the RRV, a first judgment value of the RRV corresponding to the permissible lowering limit of the arousal level is set, and the RRV at the time of awakening in the frequency band where the appearance of the respiratory fluctuation is expected, and the appearance of the respiratory fluctuation and the blood pressure fluctuation RR at waking in the frequency band where
Based on the difference between V and R, R
Set a second judgment value of RV, and set RR for the first number of samples.
Calculate V to extract respiratory fluctuations, calculate RRV for the second number of samples to extract respiratory fluctuations and blood pressure fluctuations, and compare the first judgment value with the current RRI respiratory fluctuations. ,
RRV for the first number of samples and R for the second number of samples
The blood pressure fluctuation, which is the difference from the RV, and the second determination value are compared to determine that the current arousal level has fallen below the allowable limit.

[Operation] First, the principle of respiratory fluctuation extraction in the present invention will be described with reference to FIG. 1 and FIG.

In the present invention, the frequency f ₁ near the heartbeat frequency band
When the respiratory variation of the RRI is expected to appear, the number of samples related to the RRV calculation is set so that this band is included in the pass band of the variance ratio-heart rate frequency characteristic.

In other words, considering a case where it is expected that respiratory variation in the frequency f ₁ as shown in FIG. 1 (a) appears in the present invention, first, as the number of samples n, Fig. 1 (b as shown in), the dispersion ratio band the frequency f ₁ belongs is included in the passband (passing ratio input RRV) - sample number n ₁ corresponding to the heartbeat frequency characteristic
Is set.

FIG. 1 (b) shows the variance ratio-heart rate frequency characteristic using the number n of samples as a parameter. In this figure, three types of relationships having the relationship of n ₁ <n ₂ <n ₃ (1) are shown. Variance ratio for number of samples-
The heart rate frequency characteristics are shown. As shown in this figure, the cutoff frequency increases as the number n of samples decreases.

The reason why the dispersion ratio-heart rate frequency characteristic has a high-pass characteristic and the cutoff frequency depends on the number of samples is as follows.

FIG. 2 shows the relationship between the number of samples and the RRV.

Since RRV is the variance of RRI as described above, if this is represented by a mathematical formula, Becomes Here, RRI _i is the measured value of RRI, ▲ ▼ is 1
It is the average value of ｎn RRIs.

As shown in FIG. 2 (a), when the number n of samples related to the RRV calculation is sufficiently large, the average
The value reflects the effect of the low frequency component of the I fluctuation.

Conversely, when the number of samples n is small, FIG.
As shown in (1), the low frequency component is not reflected on the average value ▲ ▼, and the low frequency component of the RRI fluctuation does not affect the deviation (▲ ▼ −RRI _i ) obtained from the average value ▲ ▼. .

Therefore, in the RRV calculation based on the equation (2), the same result as the calculation performed on the high-pass filter waveform as shown in FIG. 2C is expected.

As described above, since the average value ▼ depends on the number n of samples, the high-pass characteristic of the dispersion ratio-frequency characteristic depends on the number n of samples.

That is, the larger the number of samples n is the influence of the low frequency components of the RRI variation to reflect the RRV operation, for example, a larger number of samples n ₃ in low cut-off frequency becomes higher the smaller the number of samples n _1.

Therefore, dispersion ratio - frequency characteristic so as to include a frequency f ₁ of the respiratory variation of the RRI, set the number of samples, for example, in n _1, dispersion ratio according to the number of samples n ₁ - frequency characteristics, for example, during vehicle traveling measured RRI variation strength - heart rate (of course, may be in this reciprocal serving frequency) if Jozure the characteristic, it can be seen that detecting respiratory variation of the frequency f ₁ as shown in Figure 1 (c). This is actually achievable by operation of RRV by sample number n _1.

Based on such a principle, in claim (1) of the present invention, first, the number of samples n is set as described above. Next, based on the RRV at the time of awakening by the number of samples n,
The RRV determination value corresponding to the permissible lowering limit of the arousal level is set. Further, RRV is calculated for the number n of samples. Then, the determination value is compared with the current RRV to determine whether the current arousal level has fallen below the allowable limit.

Therefore, in claim (1) of the present invention, the number of samples n is set so that the frequency band in which respiratory fluctuation is expected is included in the pass band of the dispersion ratio-frequency characteristic. Significantly less than the number of data required for FFT analysis, n is about the cycle of respiratory fluctuations (about 3 to 5) from the principle of FIG. For this reason, when the arousal level instantaneously decreases or when the arousal level monotonously decreases, the respiratory variation is extracted in a more real-time manner.

The principle of extracting respiratory fluctuation shown in FIGS. 1 and 2 can be applied to the extraction of respiratory fluctuation and blood pressure fluctuation. Therefore, when the RRV relating to the respiratory variation and the blood pressure variation is adopted as the determination target in claim (1) of the present invention, the determination is performed based on the same principle. In this case, the case where the RRV related to the respiratory variation is set as a determination target,
Generally, the number of samples is large.

Next, the principle of extracting the blood pressure fluctuation in claim (2) of the present invention will be described with reference to FIG.

FIG. 3 shows the principle of extracting blood pressure fluctuation employed in claim (2) of the present invention.

In claim (2) of the present invention, the determination is made for both the respiratory fluctuation and the blood pressure fluctuation.

First, as shown in FIG. 3 (a), to those heartbeats strength during lowering of consciousness, it is expected to have a peak of each respiratory variation and blood pressure of variation in the frequency f ₁ and f ₂ I do. Frequency f _1, for example 0.25 [1 / be
at], both the frequency f ₂ is generally for example, 0.1 [1 / beat] has f _1> f ₂ ... relationship (3).

Such relationship corresponds to established, in claim (2) is the number of samples n ₁ and n ₂ of the set as shown in FIG. 3 (b) is performed. That is, to set the sample number n ₁ and n ₂ such that the cut-off frequency is the frequency f ₁ and f _2, respectively. Sample number n ₁ and n ₂ dispersion ratio by respectively - the heartbeat frequency characteristic has a high pass characteristic as described above (see FIG. 2).

Then, the dispersion ratio by sample number n ₂ - dispersion ratio from the heartbeat frequency characteristics with sample number n ₁ - heartbeat frequency characteristic is subtracted. The characteristic of the difference resulting from this subtraction will be
The bandpass characteristic is as shown in FIG. That is, in FIG. 3 (b), the characteristics of the difference because there is no difference in both characteristics of RRV in cardiac frequency f ₁ or more regions (n ₁₎ and RRV (n ₂₎ is the cut-off characteristic. Heart frequency slope characteristics cutoff according to the sample number n ₁ is concerned about the f ₂ or f ₁ the following areas. Number of samples n _{2 in} the region where the heartbeat frequency is f ₂ or less
Is governed by the slope of the cutoff of the characteristic according to. Therefore, the passband characteristic of the difference is the vicinity of the frequency f _2.

In view of the fact that such bandpass characteristics can be obtained by calculating the RRV difference, in claim (2), the first number of samples n ₁ and the second number of samples n ₂ are respectively shown in FIG. ) it is set in accordance with the frequency f ₁ and f ₂ of the respiratory variation and blood pressure characteristic varied as shown in.

Then, based on the RRV awake at sample number n _1,
A first determination value of the RRV corresponding to the permissible lowering limit of the arousal level is set. Furthermore, the RRV at waking in sample number n ₁
If, based on the difference between RRV, the awake at sample number n _2, a second determination of the variance of the beat between the time of the heartbeat corresponding to the reduced tolerance limit of arousal level is set.

The first and second determination values set in this way are used for the determination related to the current RRV. First, the current RRV is calculated for each of the sample numbers n ₁ and n ₂ . Furthermore, compared with the current RRV first determination value according to a sample number n _1, The second determination value, the current of the sample number n ₁
It is compared with the difference between the current RRV according to RRV and sample number n _2.

Then, when the respiratory variation and the blood pressure variation exceed the determination values, it is determined that the current arousal level has fallen below the allowable limit.

Therefore, in claim (2), a decrease in arousal level is determined based on respiratory fluctuations and blood pressure fluctuations. As a result, even if respiratory fluctuations cannot be accurately extracted due to, for example, respiratory disturbance, a decrease in arousal level is determined by extracting blood pressure-related fluctuations.

EXAMPLES Hereinafter, preferred examples of the present invention will be described with reference to the drawings.

FIG. 4 shows the configuration of an apparatus employing the arousal level determination method according to the first to third embodiments of the present invention.

In this figure, an oscillation circuit 10 oscillating at a predetermined frequency is shown.
And an ultrasonic sensor 14 for transmitting an ultrasonic wave by the output of the oscillation circuit 10 and receiving a reflected wave. The ultrasonic sensor 14 is built in a seat 16 on which the driver 12 sits,
The pulsation, or heartbeat, of the driver's 12 heart is detected by the reflection of ultrasound from the heart.

The ultrasonic sensor 14 is further connected to an amplifier 18, and the amplifier 18 is connected to a comparator 22 via a filter 20. That is, the heartbeat detected as a reflected wave by the ultrasonic sensor 14 is amplified by the amplifier 18 and
, And is converted into a pulse signal by the comparator 22.

The comparator 22 is connected to a microcomputer 24, and the microcomputer 24 is connected to the awakening means 26.

The microcomputer 24 includes an RRV according to the features of the present invention.
A device that determines arousal level by comparison.
The information relating to the heartbeat is fetched by the pulse signal from 22, and the stirring means 26 is operated according to the determination result. The stirring means 26 is operated by the driver according to a command from the microcomputer 24.
For example, a warning is issued to the driver 12 to urge the driver 12 to wake up, thereby improving the wakefulness.

FIG. 5 is a flowchart showing the operation of the microcomputer 24 in the first embodiment.

In this embodiment, after an operation start 100, first, an initial value of a heart rate is input (12).

The initial value of the heart rate refers to information on the heart rate for a predetermined number of beats at the time of awakening, and is detected by the ultrasonic sensor 14 immediately before the start of running of the vehicle, for example. The predetermined number of beats, that is, the number of samples n, is such that the frequency band in which the respiratory variation of the RRI of the driver 12 is expected, for example, 0.2 to 0.3 [1 / beat] belongs to the pass band of the dispersion ratio-frequency characteristic. It is determined. In this embodiment, 3 is adopted as the number of samples n.

That is, in the input 102 of the initial value of the heartbeat, the sixth
As shown in the figure, n ultrasonic heartbeats are detected by the ultrasonic sensor 14 with reference to the heartbeat 200 at a certain point in time, and this information is supplied to the microcomputer 24 as a pulse signal.

The microcomputer 24 calculates RRV (n) based on the input initial value of the heartbeat (104). RRV (n)
Is the RRV for the number of samples n. In this embodiment, since n = 3, the RRV calculated in step 104 is
(3). The calculation is performed according to equation (2).

Next, the determination value is set by the microcomputer 24 (106). FIG. 7 shows a concept of setting a judgment value in this embodiment.

The determination value is set by the RRV (3) calculated in step 104.
Is multiplied by a predetermined coefficient α. This coefficient α
Has a value greater than one. That is, in FIG. 7, if the determination value is set as shown by a broken line 220 with respect to the RRV (3) at the time of awakening indicated by a solid line 210,
The magnitude of the respiratory variation can be determined based on whether (3) has reached the level indicated by the broken line 220.

After the determination value is set in this way, the operation proceeds to the operation when the vehicle is running.

First, a heartbeat is detected by the ultrasonic sensor 14, and information on the heartbeat related to this detection is input to the microcomputer 24 (108).

This input is, as shown in FIG.
This is done for each heartbeat.

Further, the microcomputer 24 performs step 10
RRV (n) is calculated for the heartbeat input at 8 (110).

Thereafter, the determination value and the RRV calculated in step 110
(N) are compared (112). As described above, since the number of samples n is set to pass the respiratory variation, if the driver 12's consciousness and arousal level are reduced at this time, RRV (n) according to step 110 Is a large value. Therefore, when the RRV (n) becomes larger than a certain value, it can be considered that “the driver 12 is in a dozing state”. The above-mentioned coefficient α is “drowsy”
The determination value obtained by multiplying this coefficient α by RRV (n) at the time of awakening is the current R
RV (n) is an index as to whether or not “sleeping state” is indicated.

In the comparison 112, when the determination value is equal to or more than the current RRV (n), it can be said that the awakening degree of the driver 12 has not decreased to the point where the driver 12 can be said to be in the "sleeping state". And the above operation is repeated.

Conversely, when the current RRV (n) exceeds the determination value, the microcomputer 24 operates the awakening means 26 because it can be regarded as "sleeping state" (114).

After that, the process returns to the heartbeat input 108 again, and the operation is repeated.

In this embodiment, for example, it is possible to determine the presence or absence of the respiratory fluctuation of the RRI from the information of the heartbeat related to a small number of heartbeats, as compared with the conventional technology based on the FFT analysis. For example, about several tens of beats are obtained in the FFT analysis, whereas three beats may be used in the present embodiment.

As a result, the ability to follow the actual appearance of respiratory fluctuations is improved.

For example, consider a case where respiratory fluctuations appear instantaneously, for example, a case where the consciousness of the driver 12 instantaneously decreases. FIG. 8 shows the operation of the present embodiment and the operation of the conventional apparatus relating to the FFT analysis in this case.

In this figure, in the operation of the conventional device indicated by a broken line 400, an FFT analysis is performed based on information on a relatively large number (about several tens) of beats, and an averaged determination is performed on the beats. Therefore, a momentary decrease in consciousness cannot be detected. On the other hand, in the present embodiment, since the respiratory variation is extracted by three relatively small heartbeats, it can follow the actual respiratory variation as shown by the solid line 300 in the figure. As a result, the judgment value is represented by a two-dot chain line 50 in the figure.
If it is set to be indicated by 0, it can be determined whether the decrease in the degree of awakening has reached the dangerous level, and if the awakening means 26 is operated in the state shown by oblique lines in the figure, the degree of awakening will decrease. The danger of such vehicle running can be prevented.

Further, consider the case shown in FIG. This figure shows a comparison between the operation of the present embodiment and the operation of the present embodiment when the consciousness of the driver 12 gradually decreases, that is, when the respiratory fluctuation gradually appears.

In this figure, at the time when the respiratory fluctuation reaches the dangerous level 500, the operation 410 of the conventional apparatus has a time delay from the operation 310 of the present embodiment. This delay is caused by the difference in the number of beats to be processed.

Therefore, in this case, the present embodiment can detect the reaching of the dangerous level 500 of the respiratory variation earlier, and can operate the awakening means 26 more quickly.

FIG. 10 shows the relationship between the permissible time of consciousness reduction and the vehicle speed.

Generally, when the vehicle is running at a high speed, the consciousness reduction is allowed only for a relatively short time. That is, during a high-speed run, the “drowsy state” is not allowed to continue for a long time. For this reason, as shown in FIG. 10, the allowable time has a monotonically decreasing relationship with the vehicle speed. Specifically, since the traveling distance is proportional to the vehicle speed and the inertial force is proportional to the square of the vehicle speed, the allowable time is considered to be inversely proportional to the vehicle speed or the square of the vehicle speed.

Therefore, the allowable times a and b at the vehicle speeds V ₁ and V ₂ having a relationship of V ₁ <V _{2 have} a relationship of a> b.

In this embodiment, as described above, the followability to the respiratory variation is higher than in the past, so that it is possible to cope with a higher speed running. In the conventional apparatus, if the time delay may reflect桔坑the permissible time a in FIG. 10, as the detection-alarm "dozing state" is performed during running of the vehicle speed V _1, which can not avoid the danger is there. If the time delay in this embodiment is, for example, about the allowable time indicated by b, the vehicle speed
Driver's fast enough to also avoid the dangerous enough in V ₁
12 alarms can be issued.

In this embodiment, since the amount of arithmetic processing is smaller than that of the FF analysis, the microcomputer 24 having a smaller number of bits can be used. The operation in the microcomputer 24 can be realized by a low-frequency clock. It also simplifies the program, memory (ROM)
Can be reduced. As a result, an inexpensive microcomputer 24 can be used.

Needless to say, other heartbeat detecting means may be used instead of the ultrasonic sensor 14.

 Next, a second embodiment of the present invention will be described.

FIG. 11 shows an operation flow of the arousal level determination method according to the second embodiment of the present invention. This embodiment is realized by an apparatus having the configuration shown in FIG. 4 as in the first embodiment. FIG. 11 shows the operation of the microcomputer 24 as in FIG.

Of the operations shown in this figure, the initial steps 100-1
04 is the same as in the first embodiment. A feature of the second embodiment is that RRV (8) is calculated in addition to RRV (3), and RRVB, which is the difference between the two, is included in the determination target.

That is, the number of samples n ₁ for extracting respiratory fluctuations is 3
To the number of samples n ₂ of the extraction of the blood pressure of fluctuation is set to 8, the input from the comparator 22 obtains the each sample the number of these samples RRV (104, 116).

Here, the blood pressure fluctuation is a peak that appears in the vicinity of 0.1 [1 / beat] as the arousal level decreases as described above. When the driver 12 falls asleep due to falling asleep or the like, the blood pressure-related fluctuation appears, although not as sensitive as the respiratory fluctuation, reflecting the lowered alertness. Therefore, a decrease in the arousal level can be detected by the determination of the extraction of the blood pressure fluctuation. The respiratory fluctuation appears around 0.25 [1 / beat] as described above.
When a person has a conversation, a peak appears that confuses the respiratory fluctuation due to a decrease in the arousal level. In order to prevent erroneous recognition due to the peak, in the present embodiment, blood pressure fluctuation is used supplementarily.

As shown in FIG. 12, sample number n ₁ of the extraction of the respiratory variation is set to a smaller number than the number of samples n ₂ of the extraction of the blood pressure of fluctuation. That is, since the respiratory fluctuation appears at a higher frequency than the blood pressure fluctuation, and the cutoff frequency of the RRV characteristic (high-pass characteristic) increases as the number of samples decreases, the respiratory fluctuation and the blood pressure fluctuation respectively appear. expected frequency band is sample number n ₁ and n ₂ such that the passband is set.

In steps 104 and 116, for each of the thus set number of samples n ₁ and n _2, RRV is calculated. In this embodiment, since n ₁ = 3 and n ₂ = 8, RRV (3) and RRV (8) are calculated.

Also, the RRV calculated in steps 104 and 116
Based on (3) and RRV (8), RRVB (RRV Blood press
ure) is calculated (118). That is, RRVB is calculated based on the following equation: RRVB = RRV (8) −RRVM (3) (4) RRVM (3)
Is the average value of RRV (3) calculated from the samples used to determine RRV (8). In this case, RRV
Since the number of samples serving as the basis of the calculation of (8) is eight, and there are six types of three consecutive samples in one sample used in the calculation of the RRV (8), the calculation target of the RRVM (3) RRV (3) is six.

After step 118, determination value setting is performed (10
6).

The determination value is set by multiplying RRV (3) by a coefficient α and RRVB by a coefficient β. The coefficient α and the coefficient β are set to be larger than 1 for the same reason as in the first embodiment.

Next, the operation shifts to the operation when the vehicle is running. Also in this embodiment, Step 108 is executed as in the first embodiment.
Furthermore, for the input heart rate, RRV (3), RRV
(8) and RRVB are calculated (120).

In this embodiment, the RRV calculated in step 120
(3) and RRVB are provided for determination. That is, the determination 112
In, the RRV (3) is compared with the judgment value, and in the judgment, PPVB is compared with the judgment value.

Of these determinations, the determination value used in determination 122 is α * RRV (3) set in step 106, and the determination value used in determination 124 is β * RRVB.

That is, the determination 112 is a determination relating to the appearance of respiratory fluctuation as in the first embodiment, and when the measured value RRV (3) is larger than the determination value, it is determined that there is a possibility that the arousal level may decrease, and the determination 112 is performed.
Go to step 4; otherwise, return to step 108.

On the other hand, the determination 124 performed after the determination 112 is a determination related to blood pressure variability, and proceeds to the next step 114 only when the RRVB related to the measured value is larger than the determination value, and in other cases, Return to step 108.

That is, when the driver 12 is conscious and falls asleep or is in a state similar to this, the measured value is determined to be larger than the determination value by the determination 112, and the determination value is similarly determined to be larger than the determination value. To be judged, the following step 11
By operating the wake-up means 26 in 4, the danger due to the lowered wake-up level of the driver 12 is prevented as in the first embodiment.

Further, when a peak that is confused with the respiratory variation occurs because the driver 12 is talking, the determination 112 incorrectly determines that the peak is the respiratory variation, but the subsequent determination is made.
In 124, the measured value is not larger than the determination value, and it is determined that the blood pressure fluctuation does not appear, so that the awakening means 26 does not operate.

As a result, in the second embodiment, in addition to the effects obtained in the first embodiment, there is obtained an effect that occurrence of a malfunction of the awakening means 26 due to conversation or the like of the driver 12 is prevented.

Further, by setting the coefficients α and β, it is possible to reflect the probability that the peak appearing at the frequency of the respiratory variation is truly a respiratory variation in the determination of the arousal level. In our experiments, the appearance of a peak at the frequency of the respiratory variability was
In fact, it has been confirmed that this is not due to respiratory fluctuations, but may be due to respiratory disturbance due to conversation or the like. On the other hand, blood pressure fluctuations are less susceptible to respiratory disturbance due to conversation and the like, and if a peak occurs in a frequency region where blood pressure fluctuations are expected to occur, this may be an actual blood pressure fluctuation. high. However, blood pressure fluctuations have low sensitivity to a decrease in arousal level. Therefore, if the coefficient α is sufficiently larger than 1, the coefficient β
Is set to a value close to the comparison with 1, it is possible to quantitatively prevent erroneous detection or malfunction due to conversation or the like and to take more accurate awakening measures.

Further, in the present embodiment, if the arousal level determination is to be performed with the blood pressure variability regarded as relatively heavy, the coefficient α
And the design policy of β may be reversed from the above example. In this case, the determination is made mainly by treating the blood pressure fluctuation, which has been determined as a target in the related art, and the respiratory fluctuation. However, since FFT analysis is excluded, the effect of a quick response can be ensured.

Further, by omitting the determination value setting and the determination 112 relating to the RRV (3), it is possible to perform the agitation degree determination based only on the blood pressure fluctuation.

FIG. 13 shows the contents of the RRV calculation in the stirring degree determination method according to the third embodiment of the present invention, particularly the setting of the number of samples.

The operation and the device configuration of this embodiment are the same as those of the first embodiment, except that the number of samples n is set not according to the respiratory fluctuation but according to the blood pressure fluctuation. That is, set the number of samples n into the same value as n ₂ of the second embodiment, therefore, dispersion ratio - setting the frequency characteristic of the frequency f ₂ of the respiratory variation in the cut-off frequency.

When the sample number n is set in this way, the peak related to the determination is the peak related to the blood pressure fluctuation. Therefore, in this embodiment, a decrease in the arousal level is determined by extracting the blood pressure fluctuation, and the same effect as in the first embodiment can be obtained. In this embodiment, the respiratory variation exists in the pass band of the dispersion ratio-frequency characteristic. Therefore, even when only the respiratory fluctuation appears without the blood pressure fluctuation, it is determined that the arousal level is lowered.

[Effects of the Invention] As described above, according to the present invention, the respiratory / blood pressure fluctuation is determined and detected by the RRV, so that the followability to the actual respiratory / blood pressure fluctuation is improved. It is possible to more quickly and sensitively detect a decrease in the driver's arousal level during traveling, and to improve safety.

Furthermore, according to the present invention, since the determination of respiratory / blood pressure fluctuation is performed based on the reduction ratio of the arousal level, it is possible to eliminate the influence of the individual difference of the driver on the frequency of the respiratory blood pressure fluctuation. .

Further, according to the present invention, such effects can be realized with a simple and inexpensive device.

According to the second aspect of the present invention, even when the breathing is disturbed due to conversation or the like, the arousal level is determined based on the blood pressure fluctuation, so that the operation accuracy is improved.

[Brief description of the drawings]

FIG. 1 is a diagram showing the principle of respiratory fluctuation extraction in the arousal level determination method according to the present invention. FIG. 1 (a) is an RRI fluctuation intensity-frequency characteristic diagram in which respiratory fluctuation appears, and FIG. FIG. 1 (b) is a dispersion ratio-frequency characteristic diagram showing high-pass characteristics,
FIG. 1 (c) is an RRI fluctuation intensity-frequency characteristic diagram from which respiratory fluctuations are extracted, and FIG. 2 is a diagram showing the relationship between the variance of the interbeat time of the heartbeat and the number of samples according to the present invention. FIG. 2 (a) is a diagram showing the deviation between the interbeat time of the heartbeat when the number of samples is relatively large and the average thereof, and FIG. 2 (b) is the beat of the heartbeat when the number of samples is relatively small. FIG. 2 (c) shows a waveform after passing through a high-pass filter corresponding to FIG. 2 (b), FIG. 3 (c) shows a waveform after passing through a high-pass filter corresponding to FIG. 2 (b), and FIG. FIG. 6 is a diagram showing the principle of extracting blood pressure fluctuation in the arousal level determination method according to 2),
FIG. 3 (a) is an RRI fluctuation intensity-frequency characteristic diagram showing respiratory fluctuation and blood pressure fluctuation, FIG. 3 (b) is a dispersion ratio-frequency characteristic graph, and FIG. 3 (c) is blood pressure fluctuation. FIG. 3 (d) is a variance ratio-frequency characteristic diagram showing a band-pass characteristic which is a difference between the variance of the interbeat time of the heartbeat according to the present invention and the variance of the interbeat time of the heartbeat according to the respiratory variation. FIG. 4 is a block diagram showing the configuration of an apparatus employing the arousal level determination method according to the first to third embodiments of the present invention, and FIG. FIG. 6 is a flowchart showing the operation of the embodiment, FIG. 6 is a diagram showing the calculation target of the variance of the inter-beat time of the heartbeat in the first embodiment, and FIG. 7 is a diagram showing the concept of setting a determination value in the first embodiment. FIG. 8 is a diagram showing the operation of the first embodiment at the time of instantaneous lowering of consciousness. FIG. 9 is a diagram showing the lowering of consciousness of the first embodiment. FIG. 10 is a diagram showing an operation at the time of running, FIG. 10 is a relationship diagram of permissible consciousness reduction time-vehicle speed showing an effect of improving the followability of the first embodiment, and FIG. 11 is an awakening degree according to a second embodiment of the present invention FIG. 12 is a flowchart showing the operation of the determination method, FIG. 12 is a diagram showing the calculation target of the variance of the interbeat time of the heartbeat in the second embodiment, and FIG. 13 is the arousal level determination according to the third embodiment of the present invention. FIG. 14 is a diagram showing an example of the frequency spectrum of the RRI fluctuation intensity when the eye is closed, and FIG. 15 is a diagram showing the RRI fluctuation intensity when the eye is closed. FIG. 16 is a diagram showing another example of the frequency spectrum. FIG. 16 is a diagram showing an example of the frequency spectrum of the RRI fluctuation intensity at the time of tension. 24 …… Microcomputer ▲ ▼ …… Average value of the interbeat time of the heartbeat RRV …… RRI variance f ₂ …… Frequency of blood pressure fluctuation f ₁ …… Frequency of respiratory fluctuation n, n ₁ , n ₂ …… number of samples

──────────────────────────────────────────────────続 き Continued on the front page (58) Fields surveyed (Int. Cl. ⁶ , DB name) A61B 5/02 A61B 5/18 A61B 10/00 A61M 21/00 310 B60K 28/06

Claims (2)

(57) [Claims] The number of samples is determined so that a heart rate frequency band in which a respiratory variation or an occurrence of a respiratory variation and a blood pressure variation in a heartbeat time is expected to be included in a pass band of a variance ratio-heart rate frequency characteristic. Based on the variance of the interbeat time of the heartbeat at the time of awakening in the frequency band where respiratory fluctuations or the occurrence of respiratory fluctuations and blood pressure fluctuations are expected, Set the variance judgment value of, calculate the variance of the interbeat time of the heartbeat for the number of samples,
Extract the respiratory fluctuation or respiratory fluctuation and blood pressure fluctuation of inter-beat time of heartbeat, and compare the judgment value with the respiratory fluctuation or respiratory fluctuation and blood pressure fluctuation of current inter-beat time of heartbeat A method for determining a wakefulness level, comprising determining that the current wakefulness level has fallen below an allowable limit. 2. The method according to claim 1, wherein the first sample number is determined such that a heartbeat frequency band in which a respiratory variation in the interbeat time of a heartbeat is expected is included in a passband of a corresponding dispersion ratio-heartbeat frequency characteristic. The second sample number is determined so that the heart rate frequency band in which fluctuation and blood pressure fluctuation are expected to be included in the pass band of the corresponding variance ratio-heart rate frequency characteristic, the frequency band in which respiratory fluctuation is expected to be generated Based on the variance of the interbeat time of the heartbeat at waking, a first determination value of the variance of the interbeat time of the heartbeat corresponding to the permissible lowering limit of the arousal level is set, and a frequency band in which the appearance of respiratory fluctuation is expected Of the arousal level based on the difference between the variance of the interbeat time of the heartbeat at waking and the variance of the interbeat time of the heartbeat at waking in the frequency band where respiratory and blood pressure fluctuations are expected to appear Heart rate beats corresponding to tolerance limits A second determination value of the variance of time is set, the variance of the interbeat time of the heartbeat is calculated for the first number of samples, and the respiratory variation is extracted. The variance of the interbeat time of the heartbeat is calculated for the second number of samples. Calculating the respiratory variation and the blood pressure variation, comparing the first determination value with the respiratory variation of the current interbeat time of the heartbeat, and calculating the variance of the interbeat time of the heartbeat according to the first sample number; Comparing the blood pressure variability, which is the difference between the variance of the interbeat time of the heartbeat according to the second number of samples, and the second determination value, to determine that the current arousal level has fallen below the permissible limit. Arousal level determination method.