JPH03186713A - Abnormal driving detector - Google Patents

Abstract

PURPOSE:To give an appropriate alarm to a driver by deriving the shift of a vehicle from a correct course in a distance which the driver is gazing under running the vehicle by detection values of a radius of a curve and lateral displacement, etc., and detecting abnormal driving with high accuracy irrespective of the running state of the vehicle. CONSTITUTION:An on-vehicle image processor 16 processes an image sent from a camera 10 and calculates lateral displacement DELTAl and a yaw angle DELTAtheta of a vehicle and a curve radius R of a front running route of the vehicle. In such a way, the radius R, the lateral displacement DELTAl and the yaw angle DELTAtheta detected by the device 16 are sent to a correct steering quantity calculating means 18. On the other hand, a car speed (v) detected by a car speed sensor 12 is also inputted to the means 18. Subsequent ly, the means 18 calculates a correct steering quantity mu0 corresponding to a running state and a running route shape of the present vehicle. Next, the steering quantity mu0 is sent to an abnormality detecting means 20, and compared with the present steer ing angle mu detected by a steering angle sensor 14. In such a way, when a difference between the steering angle mu0 and mu is larger than a prescribed threshold, it is decided that abnormality such as in-sleep driving and in-loocking-aside driving, etc., and an abnormality signal is outputted to an alarm means such as a buzzer, etc.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an abnormal driving detection device, and particularly to a device that detects abnormal driving such as drowsy driving by comparing a proper steering angle with a current steering angle.

[Prior Art] Various safety devices have been proposed from the viewpoint of pursuing vehicle running safety, particularly for drowsy driving.

As an example of this type of safety device, JP-A-64-8342
There are warning devices for preventing falling asleep during automatic driving disclosed in Japanese Patent No. 3 and Japanese Unexamined Patent Publication No. 1-122734. FIG. 12 shows a plan view of this conventional device. In this conventional device, a white line 1.2 for displaying a driving lane on a road is photographed by a recognition device 5 such as a camera attached to a door mirror 4 of a vehicle 3, and a driving lane is displayed by pattern recognition from the obtained image. Recognize the white line and confirm its position.

If the position of the white line indicating the driving zone deviates from a predetermined range, the system determines that drowsy driving has occurred, and issues a warning such as a buzzer to the driver to restore normal driving. .

In other words, this device determines that drowsy driving is occurring when the vehicle begins to deviate from the white line indicating the traffic lane, and issues a warning to the driver before the vehicle deviates from the lane. Give.

[Problems to be Solved by the Invention] However, the conventional drowsy driving detection device described above has several problems. As mentioned above, JP-A-64-8
In the warning devices for preventing drowsy driving disclosed in Publication No. 3423 and Japanese Patent Application Laid-open No. 1-122734, the position of the white line indicating the driving zone on the road is confirmed from an image taken by a camera, etc. This system detects drowsy driving by detecting deviation from the white line, but if the system only makes a judgment based on deviation from the white line, it is not possible to give appropriate warnings to the driver depending on the driving condition of the vehicle. be. For example, when focusing on the vehicle speed, when the vehicle speed is high, the risk associated with drowsy driving is higher than when the vehicle speed is low, even if the deviation from the white line is the same. However, this conventional device cannot deal with such cases and has the problem of not being able to give an appropriate and effective warning.

The present invention has been developed in view of the above-mentioned problems with conventional devices, and its purpose is to detect abnormal driving such as drowsy driving or distracted driving with high accuracy regardless of the driving state of the vehicle, and to issue an appropriate warning to the driver. An object of the present invention is to provide an abnormal driving detection device for a vehicle that allows safe driving.

[Means for Solving the Problems] In order to achieve the above object, the abnormal driving detection device of the present invention includes an on-vehicle camera 10 that photographs the road ahead of the vehicle, and a vehicle-mounted camera 10 that records the traveling speed of the vehicle, as shown in FIG. Vehicle speed sensor 1 to detect
2, a steering angle sensor 14 that detects the steering angle of the vehicle, and an on-vehicle image processing device 16 that calculates the lateral displacement Δ11 yaw angle Δθ of the vehicle and the curve radius R of the road from the image obtained by the on-vehicle camera 10; Appropriateness of calculating an appropriate steering amount Llo of the vehicle with respect to the running road based on the vehicle speed v1 detected by the vehicle speed sensor 12 and the lateral displacement Δ1, yaw angle Δθ, and curve radius R calculated by the on-vehicle image processing device 16. The steering amount calculation means 18, the steering angle U detected by the steering angle sensor 14, and the calculated appropriate steering ff1u. The present invention is characterized by comprising an abnormality detecting means 20 for detecting an abnormality in the steering angle by comparing the values, and an alarm means 22 for giving a warning to the vehicle driver based on the abnormality detection signal from the abnormality detecting means 20.

[Function] The abnormal driving detection device of the present invention has such a configuration, and uses the curve radius R as the running road shape in front of the vehicle, and the vehicle speed V, lateral displacement Δl, and yaw angle Δ as the running state of the vehicle.
θ is detected by processing images obtained by an on-vehicle camera.

The appropriate steering amount calculation means calculates an appropriate steering amount according to the road shape and current driving condition from the detected value thus obtained. In other words, the distance M that the driver is looking at while the vehicle is running
The deviation of the vehicle from the appropriate course (depending on the current vehicle speed V) is determined using detected values such as the curve radius R and lateral displacement Δ1, and the amount of steering that should be done to eliminate this deviation is calculated as the appropriate steering amount. .

Then, by comparing the current steering angle from the steering angle sensor 14 and the calculated appropriate steering angle, it is calculated how much the current steering angle deviates from the appropriate steering angle. When a driver engages in abnormal driving such as falling asleep or looking inattentive, the driver's ability to recognize the road ahead is significantly reduced, and the steering angle deviates significantly from the proper steering amount that should be taken.

Therefore, by comparing the current steering angle and the calculated appropriate steering angle, it is possible to reliably determine whether or not the vehicle is operating abnormally, and it is possible to issue a warning to the driver to call attention to it.

[Example] Hereinafter, a preferred example of the abnormal operation detection device according to the present invention will be described with reference to the drawings.

First example FIG. 2 is a block diagram of the configuration of the first embodiment of the present invention.

The on-vehicle camera 10 is attached to a predetermined position of the vehicle, such as a door mirror, and photographs the road ahead of the vehicle and sends the photographed image to the on-vehicle image processing device 16. Then, the on-vehicle image processing device 16 processes the image sent from the camera 10 and calculates the lateral displacement Δl of the vehicle, the yaw angle Δθ, and the curve radius R of the road ahead of the vehicle. Various methods can be considered to calculate these physical quantities, but in this embodiment, the white line for displaying the traffic zone of the road ahead is extracted from the image taken by the camera 10, and the lateral displacement is calculated from the extracted white line. Δ
11 Yaw angle Δθ and turnip radius R are calculated. That is, first, an image signal obtained by the vehicle-mounted camera 10 is converted into digital data for each predetermined pixel. Then, a white pixel is detected from the luminance change of each pixel, and a white line for displaying a traffic zone is extracted.

Next, the tangents of the extracted white line at two points having different distances in front of the vehicle are determined, and the intersection angle of these tangents is calculated. In general, the intersection angle of tangents and the curve radius R
Since there is a certain correlation between the two, it is possible to calculate the curve radius R of the road ahead of the vehicle from the obtained intersection angle. On the other hand, the lateral displacement Δ1 of the vehicle can be calculated from the inclination of the extracted white line image and the focal length of the camera, and the yaw angle θ can be calculated from the intersection of the extracted white lines. Note that the curve radius R of the road ahead and the yaw angle Δθ of the vehicle can be determined from the image taken by such an on-vehicle camera 10.
The device that calculates
-278269.

Now, the vehicle's lateral displacement Δ1 and yaw angle Δθ of the curve radius R1 of the running road detected by the on-vehicle image processing device 16 in this manner are sent to the appropriate steering amount calculation means 18. On the other hand, the vehicle speed V detected by the vehicle speed sensor 12 composed of a magnetic resistance element or the like is also input to the appropriate steering amount calculation means 18.

The appropriate steering amount calculation means 18 includes a ROM that stores various parameters, a first-in, first-out (PIFO) memory that sequentially stores data and outputs the stored data preferentially, and a CPU that performs data calculations. The process described below is performed to calculate an appropriate steering amount uo according to the current vehicle driving condition and road shape.

That is, first, data is stored in the FIFO memory for a predetermined time T.

The error Δe is calculated from the distance ΔX traveled each time and the detected curve radius R (see Figure 3.4) and is stored sequentially.

The storage number n is the maximum vehicle speed V □ and this maximum vehicle speed V
From the forward annotation ax standard ML described below corresponding to
IIlax a+ax O .

Now, as shown in FIG.
Consider a case in which the vehicle is in a running state with a lateral displacement Δ1 away from zero and a yaw angle Δθ. The deviation of the vehicle from the appropriate course 100 at the distance that the driver would gaze at while driving the vehicle, that is, the forward gaze distance 1IitLo, is the appropriate course that would exist if it were assumed that there was no lateral displacement Δl at the current position. 100, the lateral displacement Δl that would exist if the yaw angle Δθ is 0 from the current position of the vehicle, and the deviation ε due to the yaw angle Δθ at this current position. is the sum of Therefore, this forward gaze distance. Appropriate steering ff1u must be performed at the current position in order to eliminate the deviation that would occur at ff1u. can be calculated by multiplying each of these deviations by a predetermined gain and adding them, u o −K t ε, 2Δ1+3ε0(2). Here, ε is the forward gaze distance. 10 suitable courses that would exist if you progressed by
Since it is an error from 0, it is determined using the error Δe shown in FIG. 3 or 4 as follows: ε −Σ ek −1 (3). However, in order to correspond to the forward gaze distance Lo, the maximum m that satisfies the following expression is adopted as m.

Σ　Δ1≦L.

k-1 (4) The forward gaze distance generally has a positive correlation that increases as the vehicle speed V increases, as shown in Fig. 6.
The current vehicle speed V is stored in the memory. The forward gaze distance corresponding to can be read from this ROM and the above-mentioned calculation can be performed.

Furthermore, the deviation ε0 due to the vehicle's yaw angle Δθ is the current vehicle speed V. The forward gaze distance corresponding to and yaw angle Δθ, εo=Lo・Δθ (5
Calculated by For example, To-0.5 seconds, current vehicle speed V. Let's consider a case where the speed is -30km/h.

Vehicle speed ■ or as shown in Figure 7 The locations are 5m, 10m, and 20m as shown in Figure 9.

30 m s 45 m -60 m 1...
...becomes. And the distance ΔX traveled every 0.5 seconds
The error Δe from the optimal course at is expressed as the distance traveled ΔX
When calculated from the curve radius R of the road obtained by processing the image from the in-vehicle camera 10 with the in-vehicle image processing device 16, the error in the distance Δx−5 m traveled after 0.5 seconds has elapsed from the current position, for example. Δet is the curve radius R-10
Considering that 0, 100− f (100)
2-521'/2=0.125 (m). Thereafter, the error Δe at every 0.5 seconds is determined in the same way and stored in the FIFO memory sequentially, and the data shown in FIG. 10 is obtained. And the current vehicle speed V. -30
Forward gaze distance ML corresponding to km/h. -50m is read from the ROM and this forward gaze distance ML. using ε
, is calculated. The maximum m that satisfies the above formula (4) is m
-5, so ε is 0.125+ 0.125+0+0+ 0.5133-
It becomes 0.813 (m).

In this way, ε, Δ11ε0 are calculated, and then each of these calculated values is given a predetermined gain of 1, K2,
The appropriate steering amount U is obtained by multiplying and adding.

can be found. In this embodiment, in order to appropriately modify the gains 1, K2, and K3 according to the driving characteristics of the driver, the driving characteristics of the driver are adjusted, for example, during normal driving, as shown in the block diagram of FIG. A driver model memory 22 is provided that stores the average and variance of the vehicle's lateral displacement Δl, the variance of the yaw angle Δθ, etc. at the time, and it is possible to modify the gains 1, K2, and K3 based on these values. It has become. For example, when the dispersion of the vehicle's yaw angle Δθ during normal driving is large, setting the gain of the error ε0 caused by the yaw angle Δθ at the appropriate steering amount UO to a small value of 3 will better reflect the driver's driving characteristics. It becomes possible to calculate the appropriate steering fi u o when they match.

The appropriate steering amount u for the current vehicle running state calculated by the appropriate steering amount calculation means 18 as described above is sent to the abnormality detection means 2o, and the current steering angle U detected by the steering angle sensor 14 is sent to the abnormality detection means 2o. compared to Then, the appropriate steering amount U
. When the difference between the steering angle and the current steering angle U is larger than a predetermined threshold value, it is determined that an abnormality such as drowsy driving or distracted driving has occurred, and an abnormality signal is output to an alarm means (not shown) such as a buzzer. .

Note that when the abnormality detection means 2o determines whether the current steering angle U is abnormal, the threshold value for normal light detection is appropriately modified using the data stored in the driver model memory 22 described above. Therefore, for example, when the driver is driving in a relatively rough manner, by setting the threshold value high, abnormality detection can be performed in accordance with the driving characteristics of the driver.

Second Embodiment FIG. 11 is a block diagram of a second embodiment of the present invention. A feature of the second embodiment is that a neural network is used as the appropriate steering amount calculation means. This neural network is based on the vehicle speed v1 from the vehicle speed sensor 2 and the curve radius R1 from the heavy vehicle image processing device 16.
A network unit 18 consisting of a normalization unit 18a that normalizes each detected value of lateral displacement Δ11 and yaw angle Δθ, and a formal neuron.
It is composed of b. As is well known, a formal neuron is a non-linear element with a large number of outputs, and the weighting of each human force is an element that outputs a pulse only when the sum exceeds a predetermined threshold. By appropriately setting coefficients and threshold values for weighting each input signal of this type neuron, logic gates such as an AND gate, an OR gate, and a NOT gate can be constructed. Therefore, a neural network formed by combining these formal neurons can perform various calculation processes by appropriately changing the combination. Now, in the second embodiment, the appropriate steering amount UO calculated by this neural network without using a mathematical formula is sent to the abnormality detection means 20, and the current value from the steering angle sensor 14 is sent as in the first embodiment. It is compared with the steering angle U, and when it exceeds a predetermined threshold value, it is determined that there is an abnormality and an abnormality signal is output.

In this embodiment, when the abnormality detection means 20 detects an abnormality, the proper steering ff1u is performed. Instead of comparing the difference between the current steering angle U and the current steering angle U and determining that there is an abnormality when the difference exceeds a predetermined threshold, a 17-hour counter is installed to detect the difference when the difference is above the threshold and for how long. It is also possible to configure the device to be determined to be abnormal only when a predetermined time has elapsed, and to output an abnormality signal.

Further, when the alarm generated in this embodiment is canceled by the driver, the abnormality detection means 20 determines it to be a false alarm and feeds it back to the neural network, and the weighting of each type of neuron constituting the neural network is determined by the coefficient. By modifying the threshold value, it is also possible to provide a learning function that gradually matches the driving characteristics of the driver.

[Effects of the Invention] As explained above, according to the abnormal driving detection device according to the present invention, abnormal driving such as drowsy driving and distracted driving can be detected with high accuracy and reliability, and safe driving can be achieved. It has the effect of

[Brief explanation of drawings]

FIG. 1 is a block diagram of the configuration of the abnormal operation detection device according to the present invention, FIG. 2 is a block diagram of the configuration of the first embodiment of the present invention, FIG. 3 is an explanatory diagram of the error Δe in the same embodiment, and FIG. 4 is an explanatory diagram of the FIFO memory of the same embodiment, FIG. 5 is an explanatory diagram of the appropriate steering amount in the same embodiment, and FIG. 6 is a graph diagram showing the relationship between vehicle speed V, forward gaze distance, and vL of the same embodiment. FIG. 7 is a graph showing the change in vehicle speed V in the same example. FIG. 8 is a graph showing the relationship between vehicle speed V and forward gaze distance ML in the same example. An explanatory diagram showing the detection position every 5 seconds, Fig. 10 is an explanatory diagram of the FIFO memory in the same embodiment, Fig. 11 is a block diagram of the configuration of the second embodiment according to the present invention, and Fig. 12 is the configuration of the conventional device. It is a block diagram. 0 2 4 6 8 0 Vehicle-mounted camera Vehicle speed sensor Steering angle sensor Vehicle-mounted image processing device Appropriate steering amount calculation means Abnormality detection means

Claims (1)

[Scope of Claims] An on-vehicle camera that photographs the road ahead of the vehicle; a vehicle speed sensor that detects the traveling speed of the vehicle; a steering angle sensor that detects the steering angle of the vehicle; an on-vehicle image processing device that calculates the lateral displacement, yaw angle, and curve radius of the road; and a vehicle speed detected by the vehicle speed sensor, and a lateral displacement, yaw angle, and curve radius calculated by the on-board image processing device. an appropriate steering amount calculating means for calculating an appropriate steering amount of the vehicle for the road based on the steering angle; and an abnormality for detecting an abnormality in the steering angle by comparing the steering angle detected by the steering angle sensor with the calculated appropriate steering angle. What is claimed is: 1. An abnormal driving detection device for preventing drowsy driving and inattentive driving, comprising: a detection means; and an alarm means for issuing a warning to a vehicle driver based on an abnormality detection signal from the abnormality detection means.