Apparatus and method for predicting collision

• US 8,068,134 B2
• Filed: 04/25/2006
• Issued: 11/29/2011
• Est. Priority Date: 05/13/2005
• Status: Expired due to Fees

• Current Assignee: Honda Motor Co., Ltd. (Honda Motor Company)
• Original Assignee: Honda Motor Co., Ltd. (Honda Motor Company)
• Inventors: Yoshizawa, Ryouzo
• Primary Examiner(s): Ometz; David
• Assistant Examiner(s): Quiett; Carramah J
• Application Number:
  US11/380,138
• Publication Number:
  US 20060274149A1
• Time in Patent Office: 2,044 Days
• Field of Search: 348/149, 348/116, 348/118, 348/135
• US Class Current: 348/148
• CPC Class Codes: B60R 21/0134 responsive to imminent contact with an obstacle , e.g. using radar systems
• CPC Class Codes: G01S 13/865 Combination of radar systems with lidar systems
• CPC Class Codes: G01S 13/867 Combination of radar systems with cameras
• CPC Class Codes: G01S 13/931 of land vehicles
• CPC Class Codes: G01S 2013/93271 in the front of the vehicles
• CPC Class Codes: G01S 3/20 derived by sampling signal received by an antenna system having periodically-varied orientation of directivity characteristic
• CPC Class Codes: G06V 20/58 Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads

Abstract

The present invention is a collision prediction apparatus that comprises a plurality of sensors for detecting an object in front of a vehicle by different mechanism and a control unit for selecting a sensor adequate for an input condition and predicting a collision between the vehicle and the object, based on information obtained from the selected sensor, wherein the plurality of the sensors comprise a radar for scanning front of the vehicle and a camera for taking an image of the front of the vehicle, and when a condition that it is lighter outside the vehicle than a predetermined value is input, it is enabled to configure a control unit so as to predict a collision between the vehicle and the object.

Claims

1. A collision prediction apparatus comprising:

a plurality of sensors configured to detect an object in front of a vehicle by different mechanisms, the plurality of the sensors comprising a radar configured to scan front of the vehicle and a camera configured to take an image of the front of the vehicle; and

a control unit configured to select a sensor adequate for an input condition, and based on information obtained from the selected sensor, to predict a collision between the vehicle and the object in front,

wherein in a case that a condition that the outside of the vehicle is lighter than a predetermined value, the control unit comprises;

a processing area setting mechanism configured to recognize an image of the object with the taken image of the camera and to set a processing area within the taken image to include the image of the object;

a reflection wave extraction mechanism configured to extract a reflection wave of the radar in the processing area;

a processing area object detection mechanism configured to detect an area of the object, a distance from the vehicle to the object, and a relative speed between the vehicle and the object, based on the reflection wave extracted from the processing area; and

a collision determination mechanism configured to determine a possibility of a collision between the vehicle and the object, based on the distance from the vehicle to the object and the relative speed between the vehicle and the object, and to predict the collision between the vehicle and the object, and

in a case that a condition that the outside of the vehicle is darker than the predetermined value, the control unit comprises;

an front area object detection mechanism configured to detect an area of the object, a distance from the vehicle to the object, and a relative speed between the vehicle and the object, based on a reflection wave from front of the vehicle scanned and detected by the radar; and

the collision determination mechanism configured to predict a collision between the vehicle and the object.

2. The collision prediction apparatus according to claim 1, wherein the processing area setting mechanism defines as an area of the object the area obtained by multiplying the processing area, a reflection area of an area of a reflection wave extracted from the processing area, and a reflection processing area of an area, which is estimated to be a reflection wave from the object, by weighting factors according to an input condition, based on a difference of the reflection area detected for every predetermined time;

and thereafter by adding the processing area, the reflection area, and the reflection processing area.

3. The collision prediction apparatus according to claim 1, wherein the control unit further comprises an object error detection mechanism configured to detect errors between the processing area set with the processing area setting mechanism and the area of the object detected with the processing area object detection mechanism.

4. The collision prediction apparatus according to claim 3, wherein the errors detected while the outside of the vehicle is light are added to the errors detected while the outside of the vehicle is dark to improve a prediction accuracy of the collision.

5. The collision prediction apparatus according to claim 1, wherein the processing area setting mechanism is configured to divide the processing area into a front area and side area of the object, and wherein the control unit further comprises an object movement direction detection mechanism configured to determine a possibility of a collision between the vehicle and the object, based on a movement direction of the side area.

6. The collision prediction apparatus according to claim 1, wherein the processing area setting mechanism is configured to divide the taken image at a position corresponding to a predetermined distance in front of the vehicle and to set the processing area within an area nearer than the predetermined distance.

7. The collision prediction apparatus according to claim 1, wherein the processing area object detection mechanism is configured to make an area of the object for every predetermined time, and the collision determination mechanism sequentially combines the area of the object made for every predetermined time.

8. The collision prediction apparatus according to claim 1, wherein in a relationship between a direction and intensity of the extracted reflection wave, the processing area object detection mechanism is configured to make the direction corresponding to a peak value of the intensity a direction of the object.

9. The collision prediction apparatus according to claim 8, wherein the processing area object detection mechanism is configured to make a graph consisting of a relationship between a direction and intensity of the extracted reflection wave for every predetermined time, wherein the collision determination mechanism is configured to sequentially compare the relationship between the direction and intensity of the extracted reflection wave, and based on a variation of a number of a peak value of the intensity, to determine a possibility of a collision between the vehicle and the object.

10. The collision prediction apparatus according to claim 1, wherein the processing area object detection mechanism is configured to make a graph of a waveform consisting of a relationship between a direction and intensity of the extracted reflection wave, to perform low pass filter processing to the graph, and thereafter to make the direction corresponding to a position of a gravity center of an area calculated by integrating the intensity a direction of the object.

11. The collision prediction apparatus according to claim 1, wherein in an intensity of the extracted reflection wave the processing area object detection mechanism is configured to make a graph where a plurality of peak values are connected and to make the direction corresponding to a position of a gravity center of an area calculated by integrating the intensity a direction of the object.

12. The collision prediction apparatus according to claim 1, wherein when in the processing area setting mechanism a plurality of the processing areas are set within the taken image, the control unit is configured to determine a possibility of a collision for every the processing area.

13. A collision prediction method for predicting a collision between a vehicle and an object in front, in a case that a condition that the outside of the vehicle is lighter than a predetermined value, the method comprising:

scanning front of the vehicle by radar, and based on a reflection wave of the radar, detecting direction information for indicating a direction of the object;

recognizing an image of the object from a whole of a taken image in front of the vehicle taken by camera, based on the direction information, and setting a processing area within the taken image so as to include the image of the object;

detecting an area of the object from the processing area, and based on a reflection wave of the radar extracted from the processing area,

detecting a distance from the vehicle to the object and a relative speed between the vehicle and the object; and

determining a possibility of a collision between the vehicle and the object, based on the area of the object, the distance from the vehicle to the object, and the relative speed between the vehicle and the object and

in a case that a condition that the outside of the vehicle is darker than the predetermined value, the method comprising;

detecting an area of an object, a distance from the vehicle to the object, and a relative speed between the vehicle and the object, based on a reflection wave from front of the vehicle scanned and detected by the radar; and

determining a possibility of a collision between the vehicle and the object, based on the distance from the vehicle to the object and the relative speed between the vehicle and the object, and to predict the collision between the vehicle and the object.