JP3257310B2 - Inattentive driving detection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a device for detecting unintentional driving, and more particularly to a device for detecting unintentional driving by detecting the line of sight of a vehicle driver.

[0002]

2. Description of the Related Art Hitherto, a configuration has been proposed in which a driving state of a vehicle driver is detected, an alarm is issued, and a vehicle speed is controlled. For example, JP-A-3-260900 discloses that C
The face of the vehicle driver is imaged with a CD camera, and the coordinate value of the LED attached to the center of the glasses is obtained in the image, and the coordinate value of the LED at the reference position to which the LED should be located when the driver is driving normally is obtained. A configuration is disclosed in which a face direction and an angle are obtained from a displacement of a coordinate position, and when the angle is equal to or more than an allowable angle for an allowable time or more, it is determined that the driver is driving aside and driving asleep, and an alarm is generated. Have been.

[0003]

As described above, the side-by-side driving or drowsy driving of the vehicle driver can be detected to some extent by the direction of the face or the degree of arousal, but the vehicle driver takes some thoughts during driving. A state in which the driver is driving in a random manner, that is, a so-called random driving, cannot be uniquely detected from the direction of the face. That is, during inattentive driving, the vehicle driver generally tends to gaze at a certain point in front of the driver, and the direction of the face hardly changes from that during normal driving. Because it does not change.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems of the prior art, and has as its object to provide a drunken driving detection device capable of reliably detecting a drunk driving and giving a more accurate warning. is there.

[0005]

SUMMARY OF THE INVENTION In order to achieve the above-mentioned object, a drunken driving detecting apparatus according to the present invention comprises a gaze detecting means for detecting a gaze direction of a driver, and a drama of the driver based on the detected gaze direction. have a determination means for determining degrees, further,
Environment recognition means for recognizing a traveling environment of the vehicle,
The determining means is a data used for the determination according to the driving environment.
Is characterized by changing the update time .

According to another aspect of the present invention, there is provided a drunken driving detecting apparatus according to the present invention, wherein the judging means determines the frequency of the driver in a specific gaze direction. It is characterized in that the degree of inattentiveness is determined.

According to another aspect of the present invention, the specific line of sight is at least a direction of a rearview mirror or a side mirror of the vehicle. It is characterized by being.

According to a third aspect of the present invention, there is provided a device for detecting a drunk driving, wherein the line of sight detecting means has a most frequent direction at the center of the line of sight. A case where the driver's line of sight deviates from the gazing range for a predetermined time is specified as the direction of the rearview mirror or the side mirror.

According to another aspect of the present invention, there is provided a drunken driving detecting apparatus according to the first aspect of the present invention, wherein the judging means determines a degree of dangling based on a change in a driver's line of sight. Is determined.

According to another aspect of the present invention, the change in the line-of-sight range is evaluated based on the moving distance of the line of sight. Features.

According to another aspect of the present invention, there is provided a drunken driving detecting apparatus according to the present invention, wherein the line-of-sight range change has a most frequent direction at the line-of-sight center. The predetermined gaze range is divided into a plurality of regions, and evaluation is performed based on a ratio of the number of stationary points of the line of sight in each divided region.

According to another aspect of the present invention, there is provided a drunken driving detection apparatus for detecting a dwell time of a driver's line of sight, and a drunkenness level of the driver based on the detected dwell time. have a determination means for, further,
Environment recognition means for recognizing a traveling environment of the vehicle,
The determining means is a data used for the determination according to the driving environment.
Is characterized by changing the update time .

According to another aspect of the present invention, there is provided a device for detecting a drunk driving, wherein the judging means determines a statistic of the detected stop time and a predetermined value. It is characterized in that the degree of inattentiveness is determined by comparing with a threshold value.

According to another aspect of the present invention, there is provided a drunken driving detecting apparatus according to the present invention, wherein the judging means sets the detected stopping time to a predetermined threshold value. It is characterized in that the degree of uncertainty is determined by comparing with.

According to another aspect of the present invention, there is provided a device for detecting a drunk driving, wherein the judging means comprises a plurality of thresholds. , And the degree of indiscretion is determined in a stepwise manner by comparing with each threshold value.

According to another aspect of the present invention, there is provided a drunken driving detecting apparatus comprising: a line-of-sight detecting means for detecting a direction of a driver's line of sight; and a stationary time detecting means for detecting a stationary time of the driver's line of sight. , have a judging means for judging driver's absentminded degree based on the detected viewing direction and residence time, further
Has environment recognition means for recognizing the running environment of the vehicle,
The determination means is a data used for the determination according to the traveling environment.
Data update time is changed .

[0017] To achieve the above object, aimless operation detecting apparatus of the present invention, in the aimless operation detecting apparatus according to claim 1 or claim 8 or claim 12 wherein, prior Symbol judging means, the running environment It is characterized in that the criterion for determining the degree of abstinence is changed accordingly.

[0018]

According to another aspect of the present invention, there is provided a drunken driving detecting device according to the present invention.
There is provided a reference setting means for setting a criterion based on data during normal operation, wherein the deciding means determines the degree of unskilledness using the set criterion.

Further, in order to achieve the above object, aimless operation detecting apparatus of the present invention, according to claim 1 or claim 8 or
13. The apparatus according to claim 12 , wherein the environment recognition means includes a preceding vehicle detection means.

Further, in order to achieve the above object, aimless operation detecting apparatus of the present invention, according to claim 1 or claim 8 or
13. The apparatus according to claim 12 , wherein the environment recognition unit includes a vehicle speed detection unit.

Furthermore, in order to achieve the above object, aimless operation detecting apparatus of the present invention is also Claim 1 or 8
13. The apparatus according to claim 12 , wherein the environment recognition unit includes an identification unit for identifying a traveling path.

[0023]

According to the first aspect of the present invention, the inattentive driving detection device pays particular attention to the direction of the line of sight of the vehicle driver, and determines the degree of inattentiveness based on the direction of the line of sight. As mentioned above, at the time of inattentive driving,
The driver tends to gaze at a point in front, which causes a deviation in the line of sight. Therefore, by detecting this bias, it can be determined whether or not the vehicle driver is driving indiscreetly. Here, the data of the gaze direction is statistically processed,
When judging the degree of indiscretion based on the statistical processing value,
There are differences in the degree of dispersion of individual data depending on the environment
You. For example, when driving in an urban area,
Since the data exists in the range, the variation of data becomes large.
Therefore, in order to obtain highly reliable data, the number of data must be large.
Need to be collected. Therefore, it is determined according to the driving environment
By changing the update time of the data used for
The required number of data according to the driving environment.
Determination processing can be performed.

The gaze direction can be specified by detecting the position of the iris of the eye of the vehicle driver, as described later.

According to the second aspect of the present invention, the degree of inattentiveness is determined based on the frequency of the driver in a specific gaze direction. As described above, the driver tends to gaze at a point in front of the driver when he / she is driving in a lazy manner, so that he or she does not look in the direction that the driver would normally see during normal driving, and the frequency of the direction tends to decrease. Therefore, it is possible to accurately determine the degree of indulgence based on the frequency of the driver's specific gaze direction.

According to a third aspect of the present invention, the degree of inattentiveness is determined based on at least the frequency of the direction of the rearview mirror or the side mirror of the vehicle. During normal driving, the driver looks at the rearview mirror to check the rear, or the side mirror to check the side of the vehicle.
Therefore, during normal operation, the line of sight occurs in the direction of the rear-view mirror or the side mirror at a certain frequency. On the other hand, when the driver is not driving, such rearview mirrors or side mirrors are rarely checked.
Inattentive driving can be detected based on a decrease in the frequency of the line of sight in this direction.

According to the fourth aspect of the present invention, the rearview mirror or the side mirror is checked to be out of the front gaze range during normal driving, and the predetermined gaze having the most frequent direction at the center of the line of sight is provided. The case where the driver's line of sight deviates from the range for a certain period of time is specified as the direction of the rearview mirror or the side mirror, and the line of sight to the rearview mirror or the side mirror is reliably detected.

According to the fifth aspect of the present invention, the degree of inattentiveness is determined based on a change in the driver's line of sight. In the unintentional driving detection device according to claim 2, the unintentional driving degree is determined based on the frequency of the driver's specific gaze direction. However, in light of the fact that the driver tends to gaze at a specific direction during unintentional driving, it is more directly performed. It is possible to determine the degree of indiscretion based on a change in the line of sight.

According to the sixth aspect of the present invention, the change in the line-of-sight range is evaluated based on the moving distance of the line of sight. That is, in normal driving, the line of sight is widened to widely check the environment around the vehicle, and the moving distance of the line of sight is increased. Therefore, it is possible to judge whether or not the driver is driving by evaluating the magnitude of the moving distance of the line of sight.

According to the seventh aspect of the present invention, the gaze range is divided into a plurality of regions, for example, a central region and a left and right region, and the number of sight line stopping points in each divided region is compared.
During normal operation, the gaze stop point is also dispersed in each divided area to see a wide range, but during inattentive driving there is a tendency to gaze at a certain point, so the gaze stop in a specific area, for example, the central area The points will be concentrated. Therefore, by evaluating the ratio of the number of stationary points of the line of sight in each of the divided areas, it can be determined whether or not the driver is driving casually. In addition,
“Stop” means a state in which the line of sight stays within a certain limited range for a certain period of time.

In accordance with the eighth aspect of the present invention, the inattentive driving detection device detects a stop time of the line of sight, and determines the degree of inattentiveness of the driver based on the stop time. As described above, there is a tendency that the driver gazes at a specific direction at the time of the inattentive driving, and accordingly, the stationary time of the line of sight in the specific direction tends to be longer than at the time of the normal driving. Therefore, by detecting the stop time and comparing it with a predetermined value, it is possible to reliably detect the degree of indulgence. At this point, the eye-gaze stop time data is statistically processed and
When determining the degree of indiscretion based on the statistical processing value of
There is a difference in the degree of dispersion of individual data depending on the environment
You. For example, when driving in an urban area,
Since the data exists in the range, the variation of data becomes large.
Therefore, in order to obtain highly reliable data, the number of data must be large.
Need to be collected. Therefore, it is determined according to the driving environment
By changing the update time of the data used for
The required number of data according to the driving environment.
Determination processing can be performed.

According to the present invention, the stop time is statistically processed, and the statistically processed value is compared with a predetermined threshold value. If the driver has been driving for more than a certain period of time, is the driver in a normal condition?
Or, the tendency of being in a drunken state can be determined by the statistical processing of the collected data, and by comparing the statistically processed value with a predetermined threshold value, the dispersion of the collected data can be absorbed and the tendency of long-term indiscriminate driving can be assured. Can be detected.

According to the tenth aspect of the invention,
Rather than comparing the statistical processing value of the stop time and a predetermined threshold value as in the indiscriminate driving detection device according to claim 9, the detected stop times are each compared with a predetermined threshold value,
Judgment is determined in real time. As a result, even if the stop time for one line of sight suddenly becomes long, it can be determined that the vehicle is driving insanely, and it is determined instantly in an emergency situation such as when the distance to the preceding vehicle is short. it can.

In the eleventh aspect of the present invention,
A plurality of thresholds are provided for determining whether the vehicle is in inattentive driving or not, and a statistical processing value of the stop time or each stop time is compared with each predetermined threshold. As a result, the degree of uncertainty can be evaluated stepwise, and an alarm output or the like according to the level of uneasiness can be made.

According to a twelfth aspect of the invention,
The degree of driver indecision is determined based on the gaze direction and the gaze stop time. As described above, it is possible to determine the degree of indiscretion based on either the line-of-sight direction or the stop time, but by using these two physical quantities together, it is possible to more reliably determine the degree of indiscretion of the driver. Where gaze direction
Statistical processing of the data of the time of gaze and the dwell time, and the statistical processing
When judging the degree of indecision based on the value, depending on the driving environment
Differences occur in the degree of dispersion of individual data. For example,
There are a wide range of things to watch out for in urban driving
Therefore, the variation in data becomes large. Therefore reliability
To obtain high-quality data, it is necessary to collect a large number of data.
It is necessary. Therefore, it is used for determination according to the driving environment.
By changing the data update time,
Acquiring the required number of data according to it, enabling quick judgment processing
It will work.

According to the thirteenth aspect of the invention,
When judging the degree of indiscretion based on the gaze direction and the stop time,
The criterion for the degree of indiscretion is changed in consideration of the traveling environment of the vehicle. The direction of the line of sight and the stop time of the line of sight differ depending on the city area or on the highway, and also depending on the vehicle speed and the presence or absence of a preceding vehicle.
Therefore, by changing the criterion for the degree of indiscretion according to the traveling environment, the degree of indiscretion can be reliably determined regardless of the traveling environment.

According to a fourteenth aspect of the present invention, there is provided a drunken driving detecting apparatus.
The update time of the data used for the determination of the degree of madness is changed according to the traveling environment of the vehicle. In the case where data of the gaze direction and the gaze stop time are statistically processed, and the degree of uncertainty is determined based on the statistically processed value, the degree of variation in individual data varies depending on the traveling environment. For example, when traveling in an urban area, there are a wide range of objects to be aware of, so that the variation in data becomes large. Therefore, it is necessary to collect a large number of data in order to obtain highly reliable data. Therefore, by changing the update time of the data used for determination according to the traveling environment, the required number of data according to the traveling environment can be acquired, and a quick determination process can be performed.

In the fourteenth aspect of the present invention,
When judging the degree of indifference based on the gaze direction and the dwell time of the gaze, instead of using a predetermined threshold value, the threshold value is learned during normal operation, and whether or not the vehicle is in indiscriminate driving is compared with the learned threshold value. Is determined. As a result, a threshold value can be set according to each driver, and the degree of uncertainty can be determined more adaptively.

[0039] In aimless operation detecting apparatus according to claim 15 or claim 17 wherein recognizes the preceding vehicle and the vehicle speed, a traveling road as the driving environment. When there is a preceding vehicle, the driver's driving task is smaller than when there is no preceding vehicle, so that the line-of-sight direction is relatively narrow even during normal driving, and the parking time tends to be longer. Also, the line of sight becomes narrower as the vehicle speed increases, and there is a tendency to gaze at one point. Furthermore, the range of line of sight tends to be narrower when traveling on a highway than when traveling in an urban area, and the stop time tends to be longer.

As described above, the presence / absence of the preceding vehicle, the level of the vehicle speed,
Alternatively, by changing the determination criterion according to the traveling road such as urban driving or highway driving, the driver's indiscretion can be determined more adaptively.

[0041]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows the overall configuration of this embodiment. A line-of-sight detection sensor 10 that detects the line of sight of the vehicle driver is provided near the vehicle driver's seat. This gaze detection sensor 1
As 0, for example, utilizing the fact that a strong reflection image is obtained from the eyes when the light source is placed on the line of sight, two different infrared wavelengths are irradiated to the eye of the vehicle driver, and two characteristic reflections are obtained. The gaze direction can be detected by acquiring the images and extracting the reflection image of the iris portion from the difference image, or by detecting the visible light direction as in the gaze direction detection method disclosed in Japanese Patent Application Laid-Open No. 3-165737. By comparing the difference image between the two images of the camera and the infrared light sensing camera with an appropriate threshold value, the iris portion can be extracted and the gaze direction can be detected. The line of sight detected by the line-of-sight detection sensor 10 is output to the line-of-sight counter 24 for each XY coordinate, the stop range / stop time condition determination unit 26, and the stop time within stop range measurement unit 28.

The line-of-sight counter 24 for each XY coordinate supplies the position of the line-of-sight direction from the sampling timer 36 with the line-of-sight detection range shown in FIG. 4 as the horizontal direction (X direction) and the vertical direction (Y direction) as orthogonal coordinates. It is specified at a predetermined sign pulling timing, and the number of line-of-sight points for each coordinate is counted. Thus, the count value of each coordinate point within a predetermined time is obtained. In addition, the stop range / stop time condition determination unit 26 determines whether the set of line-of-sight points has passed a predetermined time or more within a predetermined range, that is, whether or not it can be determined as a stop point. It outputs to the stop point XY coordinate-specific counter 38. When the stop range / stop time condition determination unit 26 determines that the vehicle is at a stop point, the stop time within the stop range measurement unit 28 measures the time at the stop point. If it is determined that the point is not a stop point, the data of the measurement time is discarded. The in-stoppage stop time measuring unit 28 sequentially measures the stop time in a predetermined stop range. The rear-view mirror / side-mirror position estimating unit 20 estimates a position in a line-of-sight direction that would be obtained when the vehicle driver looks at the rear-view mirror or the side mirror. Since the gaze detection sensor 10 performs gaze detection using the corneal reflection of the eye as described above, it is necessary to perform correction for each driver in consideration of the fact that the curvature of the cornea differs for each driver. It is difficult to automatically and accurately correct the absolute coordinates of the line of sight, and it is therefore necessary to estimate the position of a rearview mirror or a side mirror that differs for each driver. For this reason, in the rearview mirror / side mirror position estimating unit 20,
When confirming the rearview mirror or the side mirror, not only the line of sight but also the face move at the same time, unlike the case of gaze ahead, so that the fact that the detected gaze direction deviates greatly from the time of gaze ahead is used. FIG. 4 shows a line-of-sight detection range when the field of view during normal running is ± 10 ° in the vertical direction and 20 ° in the horizontal direction. The case where there is a large departure in a specific direction is indicated by oblique lines in the figure. Of course, the assumed range of the rear-view mirror detection and the expected range of the side mirror detection differ depending on the individual driver. Therefore, it is necessary to estimate these assumed ranges from the number of acquired data within a lapse of a predetermined time from the start of driving. The rearview mirror / side mirror check number counter 22 determines whether or not the estimated rearview mirror detection range estimated by the rearview mirror / side mirror position estimating unit 20 and the line of sight detected in the expected side mirror detection range exist. Is determined, and if there is, the count value is sequentially incremented and the number of confirmations is counted. The outputs of the stop point XY coordinate-based counter 38 and the stop time within stop range measuring section 28 are input to the rearview mirror / side mirror position estimating section 20. The count value from the rearview mirror / side mirror check number counter 22 and X
The count value from the line-of-sight counter 24 for each Y coordinate, the evaluation value from the counter 38 for each stop point XY coordinate, and the stop time value from the stop time within stop range measuring unit 28 are used as the inattentive driving determination unit 4.
It outputs the control signal to the alarm unit 42 to generate an alarm when the inattentive driving is detected.

On the other hand, a steering angle sensor 12, a vehicle speed sensor 14, and a forward monitoring sensor 16 are provided as means for recognizing the running environment of the vehicle. Note that, for example, a CCD camera or the like is used as the front monitoring sensor 16. These detection signals are stored in the driver's gaze characteristic recording unit 3 in the same manner as the detection signal from the gaze detection sensor 10 and the signal from the timer 18 described above.
0 and is stored as personal characteristics, and the driver's line of sight center coordinates are calculated. The line-of-sight center coordinates are output to the rear-view mirror / side-mirror position estimating unit 20 described above. The detection signal from the front monitoring sensor 16 is output to the preceding vehicle determination unit 32, and the presence or absence of the preceding vehicle is detected. The steering angle sensor 12 and the vehicle speed sensor 1
4. The detection signal from the forward monitoring sensor 16 is output to the traveling environment determination unit 34 to identify whether the traveling road on which the vehicle is traveling is a downtown area, an urban area, a national road (prefectural road), or an expressway. . The travel path information for identifying the travel path receives, in addition to the identification from the sensor, information indicating a road environment from an infrastructure such as a map data of a navigation system or a beacon installed on a road by radio waves or the like. A method is also acceptable. Driver gaze characteristic recording unit 3
0, the respective signals from the preceding vehicle determination unit 32, the vehicle speed sensor 14, and the traveling environment determination unit 34 are output to the inadvertent driving determination unit 40,
A threshold value is set according to the driver, and the threshold value is changed depending on the presence or absence of the preceding vehicle and the traveling environment, and the degree of uncertainty is determined. When the traveling road changes, the driving environment estimating unit 34 outputs a reset signal to the driver's gaze characteristic recording unit 30, and data is accumulated for each traveling road.

Hereinafter, the processing performed by the absent-minded driving determination unit 40 includes processing based on the line of sight, processing based on the stop time,
The process is roughly divided into a process based on the line of sight direction and the stop time.

Processing According to the Direction of the Gaze At the time of indiscreet driving, most of the consciousness is occupied by parts (thoughts) other than driving and gaze ahead. Therefore, it is expected that the rear view by the rearview mirror and the rear side monitoring by the side mirror will be sparse. FIG. 2 shows the number of confirmations (values in 10 minutes) of the rearview mirror and the side mirror during normal operation and at random operation. It can be seen that the number of confirmations of the rearview mirror and the side mirror is drastically reduced during the inattentive operation compared to the normal operation. Accordingly, for example, 60% of the number of times of confirmation of the rearview mirror and the side mirror during normal operation is set as a threshold value for determination, and by comparing the threshold value with the threshold value, it is possible to determine whether or not the vehicle is in indiscretion.

FIG. 3 is a flowchart showing a process for judging the inattentive driving based on the number of confirmations of the rearview mirror and the side mirror. In FIG. 3, first, the rearview mirror / side mirror confirmation number counter 22 counts the number of confirmations of the rear mirror and the side mirror within a predetermined time based on the detection signal from the line-of-sight detection sensor 10 (S101). This counting is performed based on whether or not the line-of-sight direction is located at the rear-view mirror position and the side-mirror position estimated by the rear-view mirror / side mirror position estimating unit 20 as described above, and the details will be described later. Then, the initial operation start (for example, 10 minutes) is learned as the number of confirmations at the time of normal operation and recorded in the driver's gaze characteristic recording unit 30 (S102). Next, the inattentive driving determination unit 40 calculates 60% (learning value × 0.6) of the count value at the time of normal driving stored in the driver's gaze characteristic recording unit 30 and sets it as a threshold. The value is compared with the latest number of confirmations for 10 minutes (S103). If the number of confirmations of the rear-view mirror and the side-mirror for the latest 10 minutes is equal to or less than the threshold value, it is determined that the driver is in an inadvertent driving state, and a control signal is output to the alarm unit 42 to generate an alarm. Do (S
104). On the other hand, if the latest number of confirmations for 10 minutes is larger than the threshold value, it is determined that the rearview mirror and the side mirror are being confirmed as in the normal operation, and no alarm is output.

As described above, by counting the number of confirmations of the rearview mirror and the side mirror and comparing the counted number with the threshold value calculated based on the number of confirmations in the normal operation, it is possible to reliably determine whether or not the vehicle is in indiscriminate driving. Can be.

In such a configuration in which the number of confirmations of the rear-view mirror and the side mirror is counted to determine whether or not the driver is inattentive, it is essential how to accurately detect the line of sight at the time of confirmation of the rear-view mirror and the side mirror. Is important. In the present embodiment, as described above, the position of the rear-view mirror and the side mirror in the line-of-sight range is estimated by the rear-view mirror / side mirror position estimating unit 20. The details are as follows.

FIG. 5 shows a processing flowchart for estimating the positions of the rear mirror and the side mirror in the rear mirror / side mirror position estimating section 20.
The basic principle of this processing flow chart is that the line of sight of the driver during normal driving is approximately ± 10 ° in the vertical direction and ± 20 ° in the horizontal direction. This is based on the fact that since the whole moves, the line-of-sight direction exceeds the detection range of the vertical, horizontal, and horizontal viewing ranges (see FIG. 4).
On the other hand, in order to check the rear and side views, it is necessary to look at the direction of the rearview mirror and the side mirror and to keep looking at the rearview mirror and the side mirror for a certain time. So, whether you checked the rearview mirror or side mirror,
Gaze deviation from a predetermined gaze detection range (± 10 ° in the upper limit direction, ± 20 ° in the left / right direction) occurs, and coordinates having a stationary time of a certain time or more and a gaze frequency of a certain number of times are rearview mirrors and side coordinates. Recognize as a mirror position. That is, in FIG. 5, it is determined whether or not the detected line-of-sight angle (line-of-sight direction) has exceeded the detection range (S201). If the detected line-of-sight angle exceeds ± 10 ° in the vertical direction and ± 20 ° in the horizontal direction, Next, it is determined whether or not it is within the expected rear-mirror detection range or the expected side-mirror detection range (S20).
2). Note that this assumed range is set in advance.
If the detected line-of-sight angle is within the range of the expected rear-mirror detection or the range of the side-mirror detection, then it is determined whether or not the line-of-sight stop time is equal to or longer than a predetermined time (for example, 0.3 seconds). (S203). This determination is made based on the output from the in-stop range stop time measuring unit 28. If the stop time is longer than a certain time, the stop point X
The number counted by the counter 38 for each Y coordinate is stored in the memory for each XY coordinate (S204), and it is determined whether or not the number of line of sight stored in the memory exceeds a predetermined number (for example, 20). (S205). If it exceeds the line of sight detection range, and has a stationary time of a fixed time or more, and occurs repeatedly, the line of sight is determined to be due to the confirmation of the rearview mirror or the side mirror, and the respective coordinates are set to the rearview mirror coordinates, It is recognized as side mirror coordinates (S206).

In this way, the rear-view mirror coordinates and the side-mirror coordinates are recognized, and the rear-view mirror / side-mirror confirmation number counter 22 determines that the detected coordinates in the line-of-sight direction coincide with these coordinates within a predetermined error range. Determines that the rearview mirror or the side mirror has been checked, sequentially increments the count value, and counts the number of checks.

In order to determine the gaze detection range shown in FIG. 4, it is necessary to extract the center of the gaze of the driver. Since the center point of the line of sight differs depending on the vehicle driver, it cannot be determined uniformly, and it is necessary to collect the line-of-sight direction data of the actual driver and extract the center point from the data.

FIG. 6 shows the operation performed by the driver's gaze characteristic recording unit 30. A processing flowchart for extracting a line-of-sight center point is shown. 6, first, based on a detection signal from the front monitoring sensor 16, the preceding vehicle determination unit 32 determines whether or not there is a preceding vehicle (S301). If a preceding vehicle is present, the steering angle is further within a predetermined angle (for example ±±) based on a detection signal from the steering angle sensor 12.
(Within 5 °) (S302), and when it is determined that the steering angle is straight ahead within the predetermined angle, the own vehicle is further determined based on the detection signal from the vehicle speed sensor 14. It is determined whether the vehicle is traveling at a predetermined speed (for example, 40 km / h) or more (S303). When the preceding vehicle exists, the vehicle goes straight ahead, and the vehicle speed is equal to or higher than a predetermined speed, the driver's line-of-sight characteristic recording unit 30 stores the line-of-sight data detected by the line-of-sight detection sensor 10 at a predetermined frequency (for example, 30 Hz). Sampling is performed and the line-of-sight coordinates are stored in a memory (S304). If the number of samples exceeds a predetermined number (for example, 10,000), the number of samples is increased to 10,000.
The coordinates having the highest frequency among the pieces of sampling data are recognized as the line-of-sight center coordinates (S305, S306).
The line-of-sight center coordinates recognized in this way are output to the rear-view mirror / side-mirror position estimating unit 20 and set as the center point of the line-of-sight detection range shown in FIG. And
The horizontal direction ± 20 ° and the vertical direction ± 10 ° are determined as the line-of-sight range (front center range).

As described above, the number of times of confirmation of the rearview mirror or the side mirror is counted based on the detected line of sight, and whether or not the driver is in a casual operation is determined by comparing with the predetermined threshold value. It is also possible to make a determination by using any one of the number of confirmations. In addition to the forward monitoring and monitoring sensor 16, a rearward monitoring sensor is also provided. It is also possible to adopt a configuration in which it is determined whether or not the vehicle is driving. It is also possible to monitor the number of confirmations in a specific direction instead of the rearview mirror or side mirror,
The point is that the fact that the frequency of a certain gaze direction changes may be used.

In the above-described embodiment, the number of confirmations of the rearview mirror and the side mirror is counted from the detected direction of the line of sight. During operation, the line of sight tends to be narrower than during normal operation. FIG. 7 shows the line-of-sight range at the time of normal operation (A) and at the time of inattentive operation (B). In the drawing, the horizontal axis and the vertical axis represent the appearance frequency at each coordinate in the horizontal and vertical line-of-sight angles (the center point of the line of sight is 0 °) in a bar graph. It can be seen that the frequency in the vicinity of the center increases during inattentive driving as compared with during normal driving, and the line of sight tends to be extremely narrow. Therefore, for example, paying attention to the horizontal viewing angle,
It is possible to determine whether or not the driver is inattentive driving based on whether or not the viewing angle is within a predetermined range.

FIG. 8 shows a processing flowchart for judging whether or not the driver is in a casual driving based on the line of sight. In FIG. 8, first, the detection signal from the visual line detection sensor 10 is stored in a memory at a predetermined frequency (for example, 30 Hz) by sampling the visual line coordinates (S401).
Specifically, the XY coordinate-based sight line counter 24 calculates the coordinates of the sight line direction based on the timing signal from the sampling timer 36, outputs the coordinates to the driver's sight line characteristic recording unit 30, and stores the same in the memory. Next, among the obtained line-of-sight coordinates,
The maximum positive and negative values of the viewing angle in the horizontal direction (X direction) are detected and stored in the memory. Then, the average value a is calculated from the absolute values of the obtained maximum positive and negative values. This calculation is performed based on the latest data for 10 minutes (S402). Next, it is determined whether or not a predetermined period of time (10 minutes at the beginning after the start of driving and 1 minute thereafter) has elapsed (S403). A traveling road (for example, a downtown area, an urban area, a national road, an expressway) is recognized, or the presence / absence of a preceding vehicle is recognized by the preceding vehicle determination unit 32 (S404), which is determined according to the recognized traveling environment. A comparison is made between the judgment value b and the average value a obtained by calculation (S405). The determination value b is set to, for example, about 15 ° on an urban road, and if the average value a is smaller than this determination value, it is determined that the vehicle is in indiscriminate driving, and an alarm is output (S406). The reason why the determination value b is changed in accordance with the traveling environment is that the line of sight changes even during normal driving according to the traveling road and the like, as described later. On national roads and highways, the line of sight during normal driving tends to be narrow.

In the processing shown in FIG. 8, it is determined whether or not the driver is inattentive driving based on the range of the horizontal line-of-sight angle. Of course, the vertical line-of-sight angle range is detected and compared with the determination value. It is also possible to determine whether or not to drive.

On the other hand, the fact that there is a change in the line-of-sight range between the inattentive driving and the normal driving as described above means that there is a difference in the moving distance of the line of sight, and the moving distance during the inattentive driving is smaller than that during the normal driving. It means that it tends to be. Therefore, it is also possible to calculate the sum of the moving distances of the line of sight (moving distances of the XY coordinate points) within a predetermined time, and to determine whether or not the driving is indiscriminate by comparing this total moving distance with the threshold value. It is possible.

FIG. 9 is a flowchart showing a process for judging whether or not the driver is in a lazy driving based on the total moving distance. In FIG. 9, first, the eye-gaze detection sensor 10
The gaze coordinates are stored in the memory at a predetermined sampling frequency based on the detection signal from (S501). Then, the moving distance of the detected line-of-sight coordinates is calculated (502). This movement distance is expressed as ((x2-y2) at the line of sight coordinates (x1, y1) at a certain sampling timing and the line of sight coordinates (x2, y2) at the next sampling timing.
x1) is calculated by ^{2 + (y2-y1) 2} ) 0.5.
Then, the calculated moving distances are sequentially added, and the latest 10
The addition sum a for one minute is calculated (S503). Next, it is determined whether or not a predetermined time (for example, 10 minutes in the initial period after the start of driving, and thereafter updated every minute) has passed (S504), and the traveling environment determination unit 34 or the preceding vehicle determination unit 32 After recognizing the driving environment (S505), a comparison is made between the set value b set based on the recognized driving environment and the addition sum a (S506). If the sum a is smaller than the set value b, it is determined that the vehicle is in the inattentive driving state, and an alarm is output (S507). The reason for changing the set value b according to the recognized driving environment is as follows.
This is because the moving distance changes in accordance with the traveling path, similarly to the determination value b in FIG. 8, and the moving distance generally tends to be shorter on an expressway than in an urban area or a downtown area.

FIG. 10 shows another processing flowchart using the moving distance. Steps S601 to S603 are the same as S501 to S503 in FIG. 9. In FIG. 10, the cumulative frequency of the stop point is further calculated, the maximum frequency b for the latest 10 minutes is calculated and stored in the memory (S604).
Then, it is determined whether a predetermined time has elapsed (S60).
5) The sum total value a of the calculated moving distances is calculated as the maximum frequency b.
Is calculated, and a magnitude comparison between this c and a set value d corresponding to the traveling environment is performed (S608). In the inattentive driving, the sum total value decreases, while the maximum frequency increases because it concentrates near the center. Therefore, the value of c = a / b becomes smaller during the inattentive operation, and the difference from the normal operation becomes clear. Therefore, the inattentive driving can be detected more reliably. In FIGS. 8 and 9, the determination is made based on the line-of-sight data with the sampling timing. However, the determination may be made based on the stop point by extracting only the stop point data from the line-of-sight data.

On the other hand, instead of calculating the total value of the moving distance and comparing it with the threshold value, the line-of-sight detection range is divided into a plurality of regions, and the detection frequency in the divided regions is counted and compared. Thus, it is also possible to determine whether or not the driver is driving casually. FIG. 11 shows an example in which the line-of-sight detection range is divided, and is divided into three regions: a central region, a left region, and a right region. Then, it is possible to determine whether or not the driver is driving indiscreetly by calculating the sum of the sight line stop points in the left region and the right region / the number of sight line stop points in the central region and comparing the calculated value with the threshold value. As described above, the line-of-sight range becomes narrower during inattentive driving, and therefore, the stop point of the line of sight tends to be concentrated in the central region as compared with the left and right regions. Therefore, the value of (left area + right area) / (center area) becomes smaller during the inattentive operation than during the normal operation, and it can be used to determine whether or not the operation is attentive.

FIG. 12 is a flow chart of a process for judging unintentional driving using the divided areas shown in FIG. First, based on the detection signal from the line-of-sight detection sensor 10, the line-of-sight coordinates of the stationary point are detected by the stationary point XY coordinate counter 38 and stored in the memory (S701). The number of stationary points located is added to calculate an addition sum a (S702). Next, the number of stationary points located in the central area is added from the line-of-sight coordinates, and the sum total b is calculated (S703). Then, it is determined whether or not a predetermined time has elapsed (S704), and a ratio c = a / b of the number of stop points in the central area to the number of stop points in the left and right areas is calculated (S705). Then, the set values d and c according to the traveling environment are compared (S706, S707), and when c is smaller than d, it is determined that the stop point is concentrated in the central area and the line-of-sight range is narrow. And output an alarm (S70
8).

Note that the stop point used in FIGS. 11 and 12 is a fixed time (for example, 0.3 seconds) within a limited range (for example, 2 × 2 °) around the start point where the line of sight is located. Is regarded as a line of sight at the same place, and this point is defined as a stop point. FIG. 13 shows an example of a stop point detected by such a definition.
During normal operation, they appear dispersed in the horizontal and vertical directions.

Stop Time In the above embodiment of the gaze direction, whether or not the driver is inattentive is determined based on the frequency and range in the specific direction. It focuses on the fact that it is different between normal driving and inattentive driving. On the other hand, there is a fact that even during the stop time of one line of sight, the normal operation is different from the inattentive operation. That is, when the driver looks at the front inattentively, the length of time per stop point of one line of sight tends to increase as compared with the time of normal operation.

In view of the above facts, an example will be described below in which it is determined whether or not the vehicle is in indiscretion based on the stop time.

FIGS. 14A and 14B are plots of the detected ratios during the entire stop time for each stop time. FIG. 14A shows the ratio during normal operation, and FIG. Is the ratio of In both figures, the ordinate indicates the dwell time every 0.1 second, and the abscissa indicates the ratio (%) in each dwell time. It can be seen that a long dwell time that does not appear during a normal operation also appears during the inattentive operation, and that the overall dwell time tends to be longer. Therefore, by monitoring the stop time, it is possible to uniquely determine whether the vehicle is operating normally or unintentionally.

FIG. 15 is a flowchart showing a process for judging whether or not the vehicle is in the inattentive operation based on the stop time. First, based on the detection signal from the visual axis detection sensor 10, the coordinates of the stop point are detected by the stop point XY coordinate counter 38 and stored in the memory (S801). Then, the stop time in the set stop range (for example, 2 × 2 に て) is measured by the stop time in stop range measuring unit 28, and the measured value is stored in the memory (S802). Then, it is determined whether or not a predetermined time has elapsed (S803). If the predetermined time has elapsed, the variance value a of the stationary time and the average value b thereof are calculated. Note that both a and b are calculated values of the latest data for 10 minutes. Then, the traveling environment determination unit 34 or the preceding vehicle determination unit 32 recognizes the traveling environment, sets a reference variance value c and a reference average value d corresponding to the traveling environment, and calculates a variance value a calculated by these. , The average value b is compared (S806, S807). And the variance a and the average b
If both are larger than the set reference value, it is determined that the pattern shown in FIG.
An alarm is output (S808).

The reason why the reference variance value and the reference average value are set in accordance with the driving environment also in this processing is that, for example, even during normal driving, the stop time tends to be shorter in urban area driving than in highway driving. This is because the variance value and the average value are considered to be smaller in the city area than in the highway.

FIG. 16 is a flow chart showing another process for determining whether or not the vehicle is in indiscriminate driving by using the stop time.
S901 to S903 correspond to S801 to S803 described above.
Then, each detected stop time is compared with the set time, and the number a and the number b shorter than the set time are calculated. The set time is set to, for example, 1.00 seconds, and is preferably changed according to the traveling environment. After calculating the long number a and the short number b, the ratio is calculated (S905), and the threshold values d and c set according to the traveling environment are compared (S907). When the number a longer than the set time is large and c is larger than d, FIG.
The pattern of FIG. 4 (B), that is, it is determined that the vehicle is in indiscreet driving, and an alarm is output (S908). In this embodiment, it is also possible to compare only a or only b with a threshold.

The processing for judging whether or not the vehicle is in a drunken operation based on the stop time has been described above. However, these processes are for judging whether or not the vehicle is in a dumb operation based on data during a predetermined time (10 minutes in this embodiment). This is to determine a long-term tendency whether the driver's situation is in a normal state or in a drunk state. On the other hand, besides outputting a warning by judging the long-term tendency of the driver as described above, it is also very effective to detect a more instantaneous dumb state and give an alarm. That is, when there is a preceding vehicle, etc., it is conceivable that an instantaneous drunken state may lead to a dangerous situation. Has a very important meaning. Therefore, in addition to the long-term driver's tendency, when the length of the stop time per one stop point exceeds a certain set threshold value, it is necessary to determine the instantaneous inattentive state.

FIGS. 17 and 18 show a process for making a more instantaneous indiscriminate driving judgment in addition to such a long-term tendency. In this process, as shown in FIG. 17, a reference stop time serving as a threshold value is determined in accordance with the level of the inattentive driving level, and an alarm is issued based on whether or not the detected instantaneous stop time exceeds the reference stop time. It is determined whether or not to give. In the present embodiment, as shown in FIG. 17, the inadvertent driving level is divided into four stages, and the reference stop time is set to 2 seconds, 3 seconds, 4 seconds, and 5 seconds for each stage. The level of the inattentive driving level is determined by providing a plurality of reference setting values (for example, four) in each processing shown in FIGS. 15 and 16, and setting each threshold value to the calculated variance value, average value, or long number. It can be evaluated stepwise by comparing the ratio of the number with the short number.

FIG. 18 is a flowchart of the process in this process. First, it is determined according to the processes shown in FIGS. 15 and 16 whether or not the driver is in a long-term indiscreet driving state and its level. (S1001). If it is determined that the driver is in a long-term inattentive driving state, the current stop time is measured by the stop time within stop range measuring unit 28 (S1002), and the measured stop time is shown in FIG. It is determined whether or not the value corresponding to the level shown in (1) is satisfied (S1003). For example, in S1001, the driver is in a long-term inadvertent driving state, and the inadvertent driving level is 1
If the measured stop time is 5 seconds,
In order to satisfy the values in FIG. 17, it is determined that an instantaneous drunken state has appeared, and an alarm is output (S1004).

As described above, the safety can be further improved by detecting the instantaneous inattentive state in addition to the tendency of long-term inattentive driving.

In the embodiment in which the line-of-sight direction and the stop time are equal to or longer than one, an example has been described in which it is determined whether or not the driver is inattentive driving based on the line-of-sight direction or the stop time. It is also possible to determine whether or not the driver is driving in a more intuitive manner.

As an example of determining using both the line-of-sight direction and the stop time, for example, as shown in FIG. 19, the front visual field range (line-of-sight detection range = see FIG. 4) is concentrically divided into a plurality of regions, Are scored so as to gradually increase as the distance approaches, and the gaze range is quantitatively evaluated based on the score. In FIG. 19, the front visual field range is divided into four regions, and the points are sequentially scored as 1, 2, 3, 4 from the outside toward the center. Then, the stop time of each stop point is measured, and the stop position of the stop point is scored according to the range of FIG.
The score is multiplied by the dwell time as a weight, and the evaluation is performed in consideration of the visual field range and the dwell time. According to this evaluation method, the evaluation value increases as the line-of-sight direction is closer to the center and as the parking time is longer, which indicates that the evaluation value increases as the degree of uncertainty increases. Therefore, it is possible to determine whether the vehicle is operating normally or inattentively based on the magnitude of the evaluation value.

FIG. 20 is a flowchart of a process based on the above concept. In FIG. 20, first, based on a detection signal from the line-of-sight detection sensor 10,
The sight line counter 38 for each XY point of the stop point calculates the coordinates of the sight line and stores the calculated coordinates in the memory of the lazy driving determination unit 40 (S11).
01). Next, the set stop range (for example, 2 × 2 ゜)
The stop time within the stop range is measured by the stop time within stop range measuring unit 28, and the stop point coordinates and the stop time at each stop point are stored in the memory (S1102). Then, based on the detected stop point coordinates and the point map for each line-of-sight range shown in FIG. 19, the points in the field-of-view coordinates are calculated (S110).
3) The evaluation value X is calculated by multiplying the score by the stop time (S1104). The evaluation values X are sequentially added (S1
At the time when a predetermined time has elapsed, the evaluation value Z per one stop point is calculated by dividing the total sum Y of the evaluation values by the total number of the stop points (S1106, S1107). Then, a comparison is made between the threshold value A set according to the driving environment and the evaluation value Z per position stop point. When the evaluation value Z is larger than the threshold value A, the gaze directions tend to converge at the center. In addition, it is determined that the vehicle is in the drunken driving state in which the stop time increases, and a warning is output (S1109, S1110).

In FIG. 19, the front field of view is divided concentrically into four regions. However, it is possible to divide the region more finely, and the number of points assigned to each region is also discontinuous from the periphery. It is also possible to set so as to increase toward the center. Also, the area division method is not concentric, and for example, as shown in FIG. 11, it is also possible to divide the area into three areas of a central area and left and right areas to obtain a score.

Further, as a method of judging unintentional driving based on the gaze direction and the stop time, for example, a logical product of the processes in FIGS. 9 and 15 is obtained, and an alarm is output only when both processes are determined to be unintentional driving. It is also possible.

As described above, the method based on the gaze direction and the stop time has been described as a method for judging whether or not the driver is in a drunk driving mode. The point is to change the threshold value when the determination is made. These are, for example, S405 in FIG.
, The processing of S506 in FIG. 9, and the processing of S806 and S807 in FIG. Here, the traveling environment includes the type of traveling road, for example, a downtown area or an urban area,
A distinction can be made between national roads and expressways, and furthermore, the presence or absence of a preceding vehicle and the level of vehicle speed can be considered. Hereinafter, the process of setting the reference value for determination according to the traveling environment will be described in detail.

FIG. 21 shows a process for setting a criterion value when the presence or absence of a preceding vehicle is considered as the driving environment.
First, the preceding vehicle determination unit 32 detects the presence or absence of a preceding vehicle based on a signal from the front monitoring sensor 16 (S1201). If there is a preceding vehicle, it is further determined whether or not it is during the personal characteristic learning period (S1202). , S1203). The personal characteristic learning period can be set to, for example, 10 minutes. When the personal characteristic learning period is set, data is collected and processed as personal characteristic values when there is a preceding vehicle. Data collection and processing, specifically the calculation of the gaze center coordinates and the number of confirmations of the rearview mirror and side mirror, the average and variance of the gaze range,
Stop time. Then, after the individual characteristic learning period ends, if there is a preceding vehicle, the reference threshold value for the determination of the inadvertent driving is changed to the case in which the preceding vehicle is present, and it is determined whether or not the driver is in indiscriminate driving (S1205). On the other hand, even when there is no preceding vehicle, it is similarly determined whether or not it is during the personal characteristic learning period (S1206). Is collected and processed (S1207), and after the individual characteristic learning period elapses, it is determined whether or not the vehicle is in indiscriminate driving using the determination reference value under the condition without the preceding vehicle (S1208).

In general, when a preceding vehicle does not exist, the driver drives while paying attention to relatively various places. However, when a preceding vehicle exists, the driver pays attention to the distance between the preceding vehicle and acceleration / deceleration. They tend to be mostly oriented and rely on the preceding vehicle for other safety checks. Therefore, the reference threshold value for the determination is changed in accordance with the presence or absence of the preceding vehicle as described above, and when the preceding vehicle is present, the reference value is relaxed as compared with the case where the preceding vehicle does not exist (specifically, By setting the threshold value of the line of sight range to be short and the threshold value of the stop time to be long), it is possible to prevent erroneous determination of inadvertent driving even though the vehicle is following the preceding vehicle normally. be able to.

As described above, when it is determined whether or not the vehicle is in indiscriminate driving only by the number of rearview mirrors, a rearward monitoring sensor is newly provided, and the rearview mirror is provided depending on whether there is a following vehicle or not. It is also conceivable to change the reference value when judging whether or not the driver is in a drive based on the number of times of confirmation. Specifically, the reference value for the judgment is increased when a following vehicle is present as compared to when the following vehicle is not present ( It is conceivable to set the number of times to be larger.

FIG. 22 shows an example in which the vehicle speed is used as the traveling environment. In FIG. 22, the horizontal axis indicates the vehicle speed, and the vertical axis indicates the rate of change of the horizontal line-of-sight angle for each vehicle speed based on 80 km / h. The higher the vehicle speed, the smaller the horizontal line-of-sight angle. It can be seen that there is a tendency to focus on one point. Therefore, for example, S4 in FIG.
The determination value b of the process in 05 is changed according to the vehicle speed,
As the vehicle speed increases, b is set to be small, and it is possible to prevent erroneous determination that the vehicle is driving insanely despite normal driving at high speed.

Although FIG. 22 shows the relationship between the vehicle speed and the horizontal line-of-sight angle change rate, the same applies to the vehicle speed and the vertical line-of-sight angle change rate. For example, in FIG. When it is determined whether or not the vehicle is driving indiscreetly using the (direction coordinates), it is conceivable to set the reference threshold value so as to decrease as the vehicle speed increases.

FIG. 23 shows another processing flowchart for changing the judgment reference value according to the vehicle speed. As described above, the reference threshold value for the determination may be increased or decreased according to the vehicle speed. However, the line-of-sight data detected by the line-of-sight detection sensor 10 is converted to the reference vehicle speed (for example, 80 k
m / h), and by comparing the data with a fixed reference threshold value, it is possible to remove the influence of the vehicle speed. That is, based on the vehicle speed data obtained by the vehicle speed sensor 14 (S1301), the data on the line of sight from the line of sight detection sensor 10 is normalized to the data of the reference vehicle speed of 80 km / h using the relationship shown in FIG. S1
302), it is determined based on the normalized data whether or not the vehicle is in indiscriminate driving (S1303).

As described above, by converting individual line-of-sight data into data at the reference vehicle speed and processing the data, it is possible to reliably detect whether the vehicle is operating normally or indiscreetly even in a driving environment in which the vehicle speed changes variously. Becomes

FIG. 24 shows an example of processing in the case where a traveling road is used as the traveling environment. In FIG. 24, the abscissa indicates four types of travel roads including a downtown area, an urban area, a national highway, and an expressway, and the ordinate indicates a rate of change based on a line of sight in an urban area. Generally, the visual field range (including the number of confirmations) increases in the order of the expressway, the national road, the city area, and the downtown area due to differences in the amount of visual stimulus, the amount of information, and the like. Therefore,
For example, by changing the determination value b in S405 in FIG. 8 according to the traveling road based on the relationship in FIG. 24, it is possible to reliably determine whether or not the driver is in a casual driving regardless of the traveling road. Note that, instead of changing the determination reference value according to the traveling path, as shown in the processing flowchart of FIG. 23, the detection data from the line-of-sight detection sensor 10 is converted into data in an urban area, and the converted data is converted into the data in an urban area. By comparing with a threshold value, it is possible to remove the influence of the traveling road.

FIG. 25 shows an example of processing in the case where a traveling path is used as the traveling environment as in FIG. 24. The horizontal axis represents the traveling path, and the vertical axis represents the stop time in the city center as the reference value. In this case, the change rate of the stop time on each traveling path in the case where the driving time is changed is shown. As in the case of FIG. 24, the stop time tends to increase in the order of a downtown area, an urban area, a national road, and an expressway based on the difference in the amount of visual stimulus and the amount of information. Thus, for example, FIG.
By changing the reference average value d in step S807 of FIG. 5 according to the traveling path in accordance with the relationship of FIG. 25, it is possible to reliably determine whether or not the driver is in a casual operation based on the stop time regardless of the traveling path. In this case as well, similarly to the case of FIG. 24, the line-of-sight data obtained by the line-of-sight detection sensor 10 is converted into city area data, and the converted data is compared with a reference time in the city area. The effect can be eliminated.

In the above, an example has been shown in which the determination reference value is changed according to the traveling environment such as the presence or absence of a preceding vehicle, the vehicle speed, and the traveling road.
Since there is a difference in the reliability of the line-of-sight data obtained by the line-of-sight detection sensor 10 depending on the traveling path, it is conceivable to change the time for starting the inattentive driving determination or the data updating time according to the traveling path. In other words, for example, the process of determining whether or not the driver is in a casual mode is not performed for 10 minutes after the start of the driving as in the process of S403 in FIG. 8, but this time may be changed according to the traveling road.

FIG. 26 shows the rate of change (based on an urban area) of the start time of the indiscreet driving determination and the data update time when the traveling environment is used as the traveling environment. This change is based on the fact that the line-of-sight data obtained in the order of downtown area, city area, national highway, and highway are less scattered and highly reliable data is easily obtained. Therefore, for example, S403 in FIG.
In the processing in the above, in the case of traveling in an urban area, it is determined whether or not the driver is indiscreet after 10 minutes from the start of driving, but it is possible to determine whether or not the driver is in a short time on an expressway, and The data update time is also updated every minute after 10 minutes in the city area,
It is also possible to determine whether or not the driver is driving in a short time with a shorter data update time.

Although FIG. 26 shows an example in which the abrupt driving determination start time and the data update time are changed in accordance with the traveling road, the individual characteristic learning period in the processing of S1203 in FIG. It is also possible to adopt a configuration in which it is changed.

FIGS. 27 and 28 show a processing flowchart in this case and a specific setting of the learning time. FIG. 27 shows how the learning time changes on each traveling road when the city area is set as a reference (10 minutes). This change is based on the fact that data obtained in the order of the downtown area, the city area, the national highway, and the highway are less scattered and reliable data is easily obtained as described above. Using this,
As shown in FIG. 28, the driving environment is recognized (S140).
1) If it is determined whether or not there is a change in the traveling environment and the traveling route changes from, for example, an urban area to a national road, the learning time is changed from 10 minutes in the urban area to about 8 minutes in the national road, The processing of S1203 and S1206 in FIG. 21 is performed (S1402, S1403, S1404).

As a result, it is possible to set an appropriate learning time according to the traveling road, and to eliminate unnecessary learning periods, thereby making it possible to quickly judge a random driving.

[0094]

As described above, according to the unintentional driving detection device of the present invention, it is possible to reliably determine whether or not unintentional driving is performed based on the driver's line of sight, and to issue an alarm or the like based on the determination result. Thereby, the safety can be further improved.

[Brief description of the drawings]

FIG. 1 is a configuration block diagram of an embodiment of the present invention.

FIG. 2 is a graph showing the number of confirmations of a rearview mirror and a side mirror during a normal operation and a lazy operation in the embodiment.

FIG. 3 is a flowchart of a random determination process based on the number of confirmations of a rearview mirror and a side mirror in the embodiment.

FIG. 4 is an explanatory diagram of a line-of-sight detection range and detection ranges of a rearview mirror and a side mirror in the embodiment.

FIG. 5 is a flowchart of a detection process of rearview mirror coordinates and side mirror coordinates in the embodiment.

FIG. 6 is a flowchart of a line-of-sight center coordinate detection process in the embodiment.

FIG. 7 is an explanatory view of a line-of-sight range at the time of normal operation and at the time of unsmooth operation in the embodiment.

FIG. 8 is a flowchart of a loose judgment process based on a line-of-sight range in the embodiment.

FIG. 9 is a flowchart of a loose judgment process based on a line-of-sight moving distance in the embodiment.

FIG. 10 is a flowchart of a loose judgment process based on a line-of-sight moving distance in the embodiment.

FIG. 11 is an explanatory diagram of area division of a visual line detection range in the embodiment.

FIG. 12 is a flowchart of a loose judgment process based on region division in the embodiment.

FIG. 13 is an explanatory diagram of a stop point in the embodiment.

FIG. 14 is a histogram diagram showing a ratio of a stop time during normal operation and during indulgent operation in the embodiment.

FIG. 15 is a flowchart of a loose judgment process based on a stop time in the embodiment.

FIG. 16 is a flowchart of a silly driving process based on the ratio of the stop time in the embodiment.

FIG. 17 is an explanatory diagram showing a relationship between a drunken driving level and a reference stop time in the embodiment.

FIG. 18 is a flowchart of a random driving determination process based on the instantaneous value of the stop time in the embodiment.

FIG. 19 is an explanatory diagram of scoring of a visual field range in the embodiment.

FIG. 20 is a flowchart of a silly driving process based on the visual field range and the stop time in the embodiment.

FIG. 21 is a flowchart of a determination reference threshold value changing process by a preceding vehicle in the embodiment.

FIG. 22 is an explanatory diagram of a change in a judgment reference value based on a vehicle speed in the embodiment.

FIG. 23 is a flowchart of a determination reference value changing process based on vehicle speed in the embodiment.

FIG. 24 is a graph showing a change in a line-of-sight range according to a traveling path in the embodiment.

FIG. 25 is a graph showing a change in a stop time depending on a traveling path in the embodiment.

FIG. 26 is a graph showing a change in a start time and a data update time of a lazy driving determination according to a traveling road in the embodiment.

FIG. 27 is a graph showing a change in a gaze characteristic learning time depending on a traveling road in the embodiment.

FIG. 28 is a flowchart of a learning time changing process based on a traveling path in the embodiment.

[Explanation of symbols]

10 eye-gaze detection sensor, 12 steering angle sensor, 14 vehicle speed sensor, 16 forward monitoring sensor, 18 timer, 20
Rearview mirror / side mirror position estimating unit, 22 Rearview mirror / side mirror check count counter, 24 X-
Gaze counter for each Y coordinate, 26 stop range / stop time condition determination unit, 28 stop time within stop range measurement unit, 30 driver gaze characteristic recording unit, 32 preceding vehicle determination unit, 34 driving environment determination unit, 36 sampling timer, 38 Stop X
-Y-coordinate-specific counter, 40 dim driving determination unit, 42 alarm unit.

Continuation of front page (56) References JP-A-63-222940 (JP, A) JP-A-63-222939 (JP, A) JP-A-62-61830 (JP, A) JP-A-60-178596 (JP, A) JP-A-59-105587 (JP, A) JP-A-56-2228 (JP, A) JP-A-7-96768 (JP, A) JP-A-6-251273 (JP, A) 6-251272 (JP, A) JP-A-6-191319 (JP, A) JP-A-6-107030 (JP, A) JP-A-5-262161 (JP, A) JP-A-5-180939 (JP, A A) JP-A-5-155269 (JP, A) JP-A-5-94600 (JP, A) JP-A-5-52608 (JP, A) JP-A-3-303933 (JP, A) JP-A-3 −260900 (JP, A) JP-A-3-165737 (JP, A) JP-A-2-30933 (JP, A) JP-A-3-24429 (JP, U) (58) Fields investigated (Int. . ^7, DB name) G01D 21/00 B60K 28/06 B60R 21/00 620

Claims (17)

(57) [Claims] 1. A possess the visual axis detecting means for detecting a viewing direction of the driver, judging means for judging driver's absentminded degree based on the detected gaze direction, and further, recognizes the traveling environment of the vehicle Environment recognition means , wherein the determination means is used for determination according to the driving environment.
A drunken driving detection device characterized by changing the update time of the data obtained . 2. The apparatus according to claim 1, wherein the determination unit determines the degree of inattentiveness based on the frequency of the driver in a specific line of sight. 3. The apparatus according to claim 2, wherein the specific gaze direction is at least a direction of a rearview mirror or a side mirror of the vehicle. 4. The lazy driving detection device according to claim 3, wherein the line-of-sight detecting means is configured to detect when the driver's line of sight deviates from a predetermined gaze range having a most frequent direction at the line-of-sight center for a predetermined time. An unintentional driving detection device characterized by specifying a direction of a mirror or a side mirror. 5. The apparatus according to claim 1, wherein the judging means judges the degree of indulgence based on a change in a driver's line of sight. 6. The unintentional driving detection device according to claim 5, wherein the change in the line of sight is evaluated based on a moving distance of the line of sight. 7. The lazy driving detection device according to claim 5, wherein the change of the line-of-sight range is performed by dividing a predetermined gaze range having a most frequent direction at the line-of-sight center into a plurality of regions, and the line-of-sight in each of the divided regions. A drunken driving detection device characterized by being evaluated based on the ratio of the number of stop points. 8. possess a residence time detecting means for detecting the residence time of the driver's line of sight, and judging means for judging driver's absentminded degree based on the detected residence time, and further, the vehicle driving environment anda environment recognizing means for recognizing, the determination means uses a decision in accordance with the running environment
A drunken driving detection device characterized by changing the update time of the data obtained . 9. The unintentional driving detection device according to claim 8, wherein the determination unit determines the degree of unintentionalness by comparing a statistical processing value of the detected stop time with a predetermined threshold value. And a lazy driving detection device. 10. The apparatus according to claim 8, wherein the determination unit determines the degree of indulgence by comparing each of the detected stop times with a predetermined threshold value. Detection device. 11. The apparatus for detecting unintentional driving according to claim 9 or 10, wherein the judging means includes a plurality of thresholds, and compares the thresholds with the thresholds to determine the degree of inadvertence in a stepwise manner. An indiscriminate driving detection device characterized by determining. 12. A gaze detecting means for detecting a gaze direction of a driver, a dwell time detecting means for detecting a dwell time of a gaze of the driver, and a degree of indulgence of the driver based on the detected gaze direction and the dwell time. possess a determination unit, and further includes an environment recognition means for recognizing a running environment of the vehicle, said-format
Means for determining the data used for the determination according to the driving environment.
A lazy driving detection device characterized by changing an update time . 13. The absent-minded driving detection device according to claim 1, wherein the judging means changes a criterion for the degree of absentmindedness in accordance with a traveling environment. apparatus. 14. The apparatus for detecting unseen driving according to claim 1, 8 or 12, further comprising: setting a criterion based on data during normal operation.
It has a constant section, the determining means, the set criterion
An inattentive driving detection device for determining the degree of inattentiveness by using the device. 15. The claim 1 or claim 8 or claim 11.
13. The device according to claim 12, wherein the environment recognizing means includes a preceding vehicle detecting means . 16. A method according to claim 1 or claim 8 or claim 7.
13. The absent-minded driving detection device according to claim 12, wherein the environment recognition unit includes a vehicle speed detection unit . 17. A method according to claim 1 or claim 8 or claim 7.
13. The absent-minded driving detection device according to claim 12, wherein the environment recognition means includes an identification means for identifying a traveling path .