JP6838391B2 - Risk estimation device - Google Patents

Description

The present invention relates to a risk estimation device that detects an obstacle and estimates the risk.

In Patent Document 1, an image captured by an in-vehicle CCD camera is captured and processed to extract forward pedestrian information in front of the vehicle, and crossing pedestrian information is captured from an external driving support road system, and the crossing is performed. A traveling support system that extracts crossing pedestrian information outside the detection range of the CCD camera from the pedestrian information and displays it on the screen as an image is disclosed.

Japanese Unexamined Patent Publication No. 2003-327065

However, depending on the driving scene, the level to be warned of pedestrians changes, and the objects to be watched are also different. However, the technique disclosed in Patent Document 1 does not take any consideration into such a point. Not.

The present invention has been devised in view of the above problems, and an object of the present invention is to provide a risk estimation device capable of estimating a risk suitable for a driving scene.

(1) In order to achieve the above object, the risk estimation device of the present invention is used for a plurality of obstacle detecting means having different monitoring distances and for each obstacle detected by the plurality of obstacle detecting means. The risk setting means for setting the risk level, the coefficient setting means for setting the weighting coefficient for each of the plurality of obstacle detection means for the risk level, and the monitoring range of the plurality of obstacle detection means. Each of the monitoring range setting means is provided with a monitoring range setting means for adjusting and setting , the coefficient setting means changes each weighting according to a driving scene, and the monitoring range setting means increases the monitoring range as the weighting coefficient increases. It is characterized by spreading.

( 2 ) Further provided with a sign or the like detecting means for detecting a road sign or a road marking existing in front of the vehicle, the coefficient setting means responds to the road sign or the road marking detected by the sign or the like detecting means. , It is preferable to change the weighting coefficient.

( 3 ) When a pedestrian crossing is detected by the sign or the like detecting means, the coefficient setting means preferably sets the weighting coefficient higher as the monitoring distance is shorter than the obstacle detecting means.

( 4 ) When the sign of the motorway is detected by the sign or the like detecting means, the coefficient setting means sets the weighting coefficient higher as the monitoring distance is longer than the obstacle detecting means. Is preferable.

( 5 ) The vehicle speed detecting means for detecting the vehicle speed of the vehicle is further provided, and the coefficient setting means increases the weighting coefficient as the monitoring distance increases as the vehicle speed increases. It is preferable to set it.

( 6 ) The steering angle detecting means is further provided, and when the sign or the like detecting means recognizes an intersection in front of the vehicle, the pedestrian crossing is recognized by the sign or the like detecting means. While there is, the weighting coefficient is set higher for the obstacle detecting means having a shorter monitoring distance, the pedestrian crossing is not recognized by the sign or the like detecting means, and the steering angle detecting means of the vehicle. When a right turn is detected, the weighting coefficient is set higher for the obstacle detecting means having a longer monitoring distance, the right turn proceeds, and the pedestrian crossing different from the pedestrian crossing is set by the sign or the like detecting means. When it is recognized, it is preferable to set the weighting coefficient higher for the obstacle detecting means having a shorter monitoring distance.

According to the present invention, a risk level is set for each obstacle detected by a plurality of obstacle detection means having different monitoring distances, and a weighting coefficient is set for each of these risk levels according to a driving scene. Since each of them is set, the monitoring distance that requires special caution is appropriately set according to the driving scene, and the degree of danger suitable for the driving scene can be estimated.

It is a control block diagram which shows the structure of the risk estimation apparatus of one Embodiment of this invention. It is a schematic plan view for demonstrating the monitoring range in the risk estimation apparatus of one Embodiment of this invention. It is a flowchart which shows an example of the control flow of the risk level estimation apparatus of one Embodiment of this invention. It is a schematic plan view for demonstrating a specific control example in the risk estimation apparatus of one Embodiment of this invention. It is a schematic plan view for demonstrating a specific control example in the risk estimation apparatus of one Embodiment of this invention. It is a schematic plan view for demonstrating a specific control example in the risk estimation apparatus of one Embodiment of this invention.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. It should be noted that the embodiments shown below are merely examples, and there is no intention of excluding the application of various modifications and techniques not specified in the following embodiments. The configurations of the following embodiments can be variously modified and implemented without departing from their gist.
An embodiment of the present invention will be described with reference to FIGS. 1 to 6. In the following explanation, the direction of travel of the vehicle is set to the front, and the left and right are determined with reference to the front. In addition, both the vehicle width direction and the road width direction are called the width direction.

[1. Constitution]
As shown in FIG. 1, the risk estimation device 20 of the present embodiment includes a short-distance sensor (obstacle detecting means) 1, an in-vehicle camera (obstacle detecting means) 2, and a long-distance sensor (obstacle detection means). Means) 3, a vehicle speed sensor (vehicle speed detecting means) 4, a steering angle sensor (steering angle detecting means) 5, and functional elements 11 to 14 of the ECU 10. Specifically, the functional elements 11 to 14 include a sign or the like detection unit (mark or the like detection means) 11, a risk level setting unit (risk level setting means) 12, a coefficient setting unit (coefficient setting means) 13, and a monitoring range setting unit. (Monitoring range setting means) 14. The ECU 10 is an ordinary computer, and internally includes a well-known CPU, ROM, RAM, I / O, and a bus line for connecting these configurations.

Both the short-distance sensor 1 and the long-distance sensor 3 detect an obstacle in front by using electromagnetic waves or the like, and in this embodiment, they are configured by radar (short-distance radar, long-distance radar). .. The in-vehicle camera 2 also has a function of detecting an obstacle in front of the vehicle. Hereinafter, when the short-distance sensor 1, the in-vehicle camera 2 and the long-distance sensor 3 are collectively referred to, they are referred to as monitoring devices 1, 2, and 3.
The monitoring ranges R1, R2, and R3 of the monitoring devices 1, 2, and 3 are usually as shown in FIG. 2, and the monitoring is the distance of the target that is supposed to be monitored by the monitoring devices 1, 2, and 3. The distances L1, L2, and L3 are set longer in this order.
Further, the monitoring width has a characteristic that is inversely proportional to the monitoring distance, and the monitoring widths W1, W2, and W3 of the monitoring devices 1, 2, and 3 are narrowed in this order.
Further, the monitoring ranges R1, R2, and R3 of the monitoring devices 1, 2, and 3 can be adjusted, respectively.

The in-vehicle camera 2 is provided at the front end of the own vehicle, the upper part of the vehicle interior, or the like, captures an image of the surroundings in front of the vehicle (hereinafter referred to as "vehicle front image"), and acquires vehicle front image data (image information). ) Is output to the ECU 10. Further, the in-vehicle camera 2 is composed of a stereo camera, and can detect the positional relationship between the own vehicle and an obstacle based on the image in front of the vehicle.
The vehicle speed sensor 4 detects the traveling speed of the own vehicle based on the rotation speed of the axle.
The steering angle sensor 5 detects the steering angle of the steering wheel.

The sign detection unit 11 analyzes the vehicle front image captured by the in-vehicle camera 2 and performs template matching with the image information of the road sign or road marking stored in advance by performing well-known image analysis of the road sign or road marking. Detect the type.

The risk level setting unit 12 analyzes the reflected wave from the sensors 1 and 3 and the front image captured by the in-vehicle camera 2, and sets the risk level D in which the risk level is quantified. It should be noted that the risk level D can be said to be a caution level that should be watched when the own vehicle travels, and can also be said to be a level at which safety measures are required.
Specifically, the risk setting unit 12 detects an obstacle based on the detection results (reflected wave or forward image) of the monitoring devices 1, 2 and 3, and estimates and detects the type and size of the obstacle. Depending on the type and size of the obstacle, the risk level d1 corresponding to the detection result of the short-distance sensor 1, the risk level d2 corresponding to the detection result of the in-vehicle camera 2, and the detection result of the long-distance sensor 3 Set the corresponding risk level d3. Then, the risk setting unit 12 uses the weighting coefficient k1 corresponding to the short-distance sensor 1, the weighting coefficient k2 corresponding to the in-vehicle camera 2, and the weighting coefficient k3 corresponding to the long-distance sensor 1 by the following equation (1). ) Sets the risk level D. The weighting coefficients k1, k2, and k3 will be described later.
D = k1 × d1 + k2 × d2 + k3 × d3 ... (1)
When this risk level D exceeds the threshold value, various controls for ensuring safety are performed, for example, a notification is given to call attention to the driver, or a brake prefill is activated to brake. Preparations are made so that the braking force can be exerted immediately in response to the driver's braking operation.
In addition, instead of the case where the risk level D exceeds the threshold value, any one of the terms (“k1 × d1”, “k2 × d2” and “k3 × d3”) of the above equation (1) exceeds the threshold value. In some cases, various controls may be performed to ensure safety.

The coefficient setting unit 13 appropriately sets the weighting coefficients k1, k2, and k3 based on the detection result of the sign and the like detection unit 11.
Specifically, the coefficient setting unit 13 sets each weighting coefficient k1 in standard time, that is, when a specific road sign or road sign is not detected by the sign or the like detection unit 11, or when the vehicle speed V does not exceed the threshold value V0. , K2, k3 are set with standard values of the same value, for example, each is set to 1.0. On the other hand, the coefficient setting unit 13 determines when a specific road sign (for example, a sign indicating a motorway or a sign indicating a pedestrian crossing) or a specific road marking (for example, a display indicating a pedestrian crossing) is detected, or the vehicle speed V. In the case of specific driving scenes such as when driving on a motorway or highway, or driving near a pedestrian crossing, such as when the threshold value V0 is exceeded, these weighting coefficients are adjusted according to each driving scene. K1, k2, and k3 are set to different values, and different weights are applied to the detection results of the monitoring devices 1, 2, and 3.
This is mainly due to the consideration that whether the obstacle to be watched should be a short-distance obstacle or a distant obstacle from the own vehicle 50A changes depending on the driving scene.
How to specifically set the weighting coefficients k1, k2, and k3 according to what kind of driving scene will be described later.

The monitoring range setting unit 14 sets the monitoring ranges R1, R2, and R3 of the monitoring devices 1, 2, and 3, respectively, and in the present embodiment, particularly when the weighting coefficient k1 for the short-distance sensor 1 is larger than the standard value. That is, when the appearance of an obstacle at a short distance from the own vehicle 50A should be warned. The monitoring range R1 of the short-distance sensor 1 is set to have a wider monitoring width W1 and a shorter monitoring distance L1 than in the normal state shown in FIG.

[2. Control flow]
Hereinafter, the control flow of the risk estimation device 20 as an embodiment of the present invention will be described with reference to FIG. This control flow is repeatedly executed at a predetermined cycle.
First, in step S10, the detection results are acquired from the monitoring devices 1, 2, and 3.
Next, in step S20, the risk levels d1, d2, and d3 corresponding to the detection results of the monitoring devices 1, 2, and 3 are calculated by the risk level setting unit 12, respectively.
Then, in step S30, it is determined whether or not a specific driving scene has been detected, and if a specific driving scene is detected, the process proceeds to step S40, and if not, the process proceeds to step S50.
In step S40, a weighting coefficient is set by the coefficient setting unit 13 according to the specific driving scene detected in step S30, the risk level D is calculated by the risk level setting unit 12 accordingly, and the monitoring range is further increased. The monitoring range is set by the setting unit 14.
In step S50, the weighting coefficients k1, k2, and k3 are set to standard values by the coefficient setting unit 13 for each of the monitoring devices 1, 2, and 3, and the monitoring ranges R1, R2, and R3 are set by the monitoring range setting unit 14. Set to the standard range.

[3. Concrete example]
Hereinafter, specific examples of changing the weighting coefficients k1, k2, and k3 according to the driving scene will be described with reference to FIGS. 4 to 6.

[3-1. Specific example 1]
In the example shown in FIG. 4, the own vehicle 50A is traveling on a general road, and the sign detection unit 11 detects a road sign (not shown) indicating the pedestrian crossing 100 in front at the position indicated by the alternate long and short dash line. is there. In FIG. 4, only the monitoring range R1 of the short-distance sensor 1 is shown for convenience at the position indicated by the alternate long and short dash line.
The coefficient setting unit 13 and the monitoring range setting unit 14 assume that there is a risk of pedestrians jumping out at a short distance from the own vehicle 50A in response to the detection of the road sign of the pedestrian crossing by the sign detection unit 11. By performing the following controls, the alertness for short-distance obstacles is increased.

That is, the weighting coefficient k1 corresponding to the detection result of the short-distance sensor 1 is set to a value larger than the standard value. That is, the weighting coefficient k1 of the short-distance sensor 1 is set higher than the weighting coefficients k2 and k3 (for example, k2 = 1, k3 = 1) of the in-vehicle camera 2 and the long-distance sensor 3 which are set to the standard values, respectively. (For example, k3 = 4), the contribution of the detection result of the short-distance sensor 1 in the calculation of the risk D by the above equation (1) is increased.
In FIG. 4, the monitoring range setting unit 14 monitors the monitoring range of the short-distance sensor 1 with a solid line emphasizing short-distance monitoring from the own vehicle 50A from the monitoring range R1 which is the standard range indicated by the alternate long and short dash line. Change to the range R1'. That is, the monitoring distance L1'of the monitoring range R1'is set closer than the monitoring distance L1 of the monitoring range R1, and the monitoring width W1'of the monitoring range R1'is set wider than the monitoring width W1 of the monitoring range R1. ..
Hereinafter, such control between the coefficient setting unit 13 and the monitoring range setting unit 14 is referred to as “pop-out warning control”.
Then, when the own vehicle 50A passes through the pedestrian crossing 100 and the sign or the like detection unit 11 does not detect the pedestrian crossing 100, the pop-out warning control is canceled and the coefficient setting unit 13 returns the weighting coefficient k1 to the standard value. , The monitoring range setting unit 14 returns the monitoring range to the monitoring range R1 which is the standard range.

Instead of the case where the road sign of the pedestrian crossing is detected by the sign or the like detection unit 11, the above control may be started when the pedestrian crossing 100 is detected by the sign or the like detection unit 11.

[3-2. Specific example 2]
The example shown in FIG. 5 is an example in which the own vehicle 50A is traveling on a high-speed motorway.
The coefficient setting unit 13 sets a larger weighting coefficient k3 corresponding to the detection result of the long-distance sensor 3 as the vehicle speed V detected by the vehicle speed sensor 4 increases. In the present embodiment, the vehicle speed V is larger than the threshold value V0. When it becomes high, the weighting coefficient k3 is set to a value larger than the standard value. Further, the coefficient setting unit 13 sets the weighting coefficients k1 and k2 of the short-distance sensor 1 and the in-vehicle camera 2 to be smaller than the standard values.
On highways, pedestrians do not jump out from the side in principle, so by setting the weighting coefficient in this way, the long-distance sensor 3 is detected when calculating the risk level D in the above equation (1). In order to increase the contribution of the result, the alert for obstacles at a long distance from the own vehicle 50A such as the preceding vehicle 50B is increased.
Instead of the case where the vehicle speed V becomes higher than the threshold value V0, the above control may be started when the sign or the like detection unit 11 detects the road sign of the motorway.

[3-3. Specific example 3]
The example shown in FIG. 6 is an example in which the own vehicle 50A is about to turn right at an intersection 200 on a general road.
When the sign or the like detection unit 11 detects a road sign (not shown) indicating the intersection 200 or detects a pedestrian crossing 100A in front of the intersection 200, the above-mentioned pop-out warning control described in Specific Example 1 is started. That is, the coefficient setting unit 13 increases the weighting coefficient k1 for the short-distance sensor 1, and the monitoring range setting unit 14 expands the monitoring range of the short-range sensor 1.
Then, in the state shown in FIG. 6, when the sign or the like detection unit 11 does not detect the pedestrian crossing 100A and the right turn of the own vehicle 50A is detected based on the steering angle sensor 5, the vehicle moves forward from the pop-out warning control. The control is changed to warn the oncoming vehicle 50C (hereinafter referred to as "oncoming vehicle warning control"). That is, the weighting coefficient k1 of the short-distance sensor 1 is returned to the standard value, the coefficient setting unit 13 makes the weighting coefficient k3 of the long-distance sensor 3 larger than the standard value, and the weighting coefficient k3 of the long-distance sensor 3 is set to the short-distance sensor. It is made larger than the weighting coefficient k2 of 1 and the in-vehicle camera 2.
Then, when the right turn of the own vehicle 50A proceeds from the state shown in FIG. 6 and the sign detection unit 11 detects another pedestrian crossing 100B, the oncoming vehicle warning control is switched to the jumping warning control, and the own vehicle 50A crosses the pedestrian crossing 100B. Jump-out warning control is performed until the pedestrian crossing 100B is no longer detected after passing.

[4. effect]
According to one embodiment of the present invention, risk levels d1, d2, d3 are set for each obstacle detected by the monitoring devices 1, 2, and 3, but these risk levels d1, d2, d3 are set. The weighting coefficients k1, k2, and k3 are set according to the traveling scene. For example, in a driving scene where caution is required for a short distance from the own vehicle 50A, the weighting coefficient k1 of the short-distance sensor is set higher than the weighting coefficients k2 and k3 of the in-vehicle camera 2 and the long-distance sensor 3. In a driving scene where caution is required for a long distance from the own vehicle 50A, the weighting coefficient k3 of the long-distance sensor 3 is set higher than the weighting coefficients k1 and k2 of the short-distance sensor 1 and the in-vehicle camera 2. To.
Therefore, it is possible to provide driving support suitable for the driving scene.

[5. Others]
(1) The risk level setting unit 12 may set the risk level d1, d2, d3 higher as the rate of increase of the weighting coefficients k1, k2, k3 with respect to time increases. For example, when the weighting coefficient k3 of the long-distance sensor 3 is increased as the vehicle speed V increases, the risk level d3 corresponding to the obstacle detected by the long-distance sensor 3 is set as the rate of increase of the weighting coefficient k3. Increase accordingly.

(2) The method of setting the weighting coefficient is not limited to the above embodiment, and the weighting coefficient of the monitoring device that monitors the main warning target may be relatively higher than the weighting coefficient of the other monitoring device. For example, if the monitoring device that monitors the main warning target is a long-distance sensor 3, the weighting coefficient k3 may be higher than the weighting coefficients k1 and k2, and the weighting coefficients k1 and k2 may be higher than the weighting coefficient k3. May be lower, or the weighting coefficients k3 may be made higher than the standard value and the weighting coefficients k1 and k2 may be made lower than the standard value.

(3) In the above-described embodiment, the road sign or road marking is detected by the sign or the like detecting unit 11, but the own vehicle position detecting means and the map information acquiring means may be used instead of the sign or the like detecting unit 11. .. That is, by comparing the position of the own vehicle with the map information, it may be detected that there is a road sign or a road marking in front of the vehicle or that the vehicle is traveling on a highway or a motorway.

(4) Further, the present invention is described in [3. The application is not limited to the driving scenes listed in the column of [Specific example], and can be applied to, for example, a driving scene traveling in the vicinity of a traffic light.

(5) In the above-described embodiment, the example in which the present invention is applied to the monitoring of the front of the vehicle has been described, but the present invention may be applied to the monitoring of the side of the vehicle or the rear of the vehicle.

1 Short-distance sensor (obstacle detection means)
2 In-vehicle camera (obstacle detection means)
3 Long-distance sensor (obstacle detection means)
4 Vehicle speed sensor (vehicle speed detection means)
5 Steering angle sensor (steering angle detecting means)
10 ECU
11 Signs, etc. detection unit (Signs, etc. detection unit)
12 Danger level setting unit (risk level setting means)
13 Coefficient setting unit (coefficient setting means)
50A Own vehicle 50B Preceding vehicle 50C Oncoming vehicle 100, 100A, 100B Pedestrian crossing 200 Intersection R1, R2, R3 Monitoring range k1, k2, k3 Weighting coefficient d1, d2, d3 Danger

Claims (6)

Multiple obstacle detection means with different monitoring distances,
A risk setting means for setting a risk level for each obstacle detected by the plurality of obstacle detection means, and a risk setting means.
A coefficient setting means for setting a weighting coefficient for each of the plurality of obstacle detection means with respect to the risk level, and a coefficient setting means .
It is provided with a monitoring range setting means for adjusting and setting the monitoring range of each of the plurality of obstacle detection means.
The coefficient setting means changes each of the weights according to the driving scene, and changes the weighting .
The monitoring range setting means is a risk estimation device, characterized in that the higher the weighting coefficient, the wider the monitoring range . Further equipped with a detection means such as a road sign or a sign for detecting a road marking existing in front of the vehicle,
The risk estimation device according to claim 1 , wherein the coefficient setting means changes the weighting coefficient according to the road sign or the road marking detected by the sign or the like detecting means. When a pedestrian crossing is detected by the sign or the like detecting means, the coefficient setting means sets the weighting coefficient higher as the monitoring distance is shorter than the obstacle detecting means. Item 2. The risk estimation device according to item 2. The coefficient setting means is characterized in that when a sign on a motorway is detected by the sign or the like detecting means, the weighting coefficient is set higher as the monitoring distance is longer than that of the obstacle detecting means. The risk estimation device according to claim 2 or 3. Further provided with a vehicle speed detecting means for detecting the vehicle speed of the vehicle,
Any one of claims 2 to 4 , wherein the coefficient setting means sets the weighting coefficient higher as the vehicle speed increases and the obstacle detecting means has a longer monitoring distance. The risk estimation device described in the section. Further equipped with steering angle detecting means,
The coefficient setting means is
When an intersection is recognized in front of the vehicle by the sign or the like detecting means,
While the pedestrian crossing is recognized by the sign or the like detecting means, the weighting coefficient is set higher for the obstacle detecting means having a shorter monitoring distance.
When the pedestrian crossing is not recognized by the sign or the like detecting means and the right turn of the vehicle is detected by the steering angle detecting means, the weighting coefficient is increased as the monitoring distance is longer than the obstacle detecting means. Set,
When the right turn progresses and the pedestrian crossing other than the pedestrian crossing is recognized by the sign or the like detecting means, the weighting coefficient is set higher for the obstacle detecting means having a shorter monitoring distance. The risk estimation device according to any one of claims 2 to 5, wherein the risk estimation device is characterized.

Images