Apparatus and method for deciding a maneuver of a nearby vehicle

Patent metadata

Field	Value
Publication number	US-10647316-B2
Application number	US-201815971500-A
Country	US
Kind code	B2
Filing date	May 4, 2018
Priority date	Dec 19, 2017
Publication date	May 12, 2020
Grant date	May 12, 2020

Abstract

An apparatus and a method are capable of accurately deciding a maneuver of a surrounding vehicle. The apparatus includes a first surrounding vehicle information detector configured to obtain first surrounding vehicle information for a surrounding vehicle of a vehicle by using a front radar device. The apparatus further includes a second surrounding vehicle information detector configured to obtain second surrounding vehicle information for the surrounding vehicle by using a corner radar device. The apparatus also includes a processor configured to decide a motion of the surrounding vehicle by using the first surrounding vehicle information and the second surrounding vehicle information. The processor is also configured to decide a maneuver of the surrounding vehicle by using maneuver decision logic derived by a mechanical training technique. The processor is further configured to decide a final maneuver of the surrounding vehicle by using the two decision results.

First claim

What is claimed is:

1. An apparatus for deciding a maneuver of a surrounding vehicle, the apparatus comprising:
a first surrounding vehicle information detector configured to obtain first surrounding vehicle information for a first surrounding vehicle relative to a vehicle-by using a front radar device;
a second surrounding vehicle information detector configured to obtain second surrounding vehicle information for a second surrounding vehicle by using a corner radar device; and
a processor configured to set a gate based on a location of the first surrounding vehicle, identify whether the second surrounding vehicle is located within the set gate, fuse the first surrounding vehicle information and the second surrounding vehicle information when the second surrounding vehicle is located within the set gate, decide a motion of the surrounding vehicle by using the first surrounding vehicle information and the second surrounding vehicle information, decide a maneuver of the surrounding vehicle by using maneuver decision logic derived by a machine training technique based on the fused surrounding vehicle information, and decide a final maneuver of the surrounding vehicle by using the two decision results.

2. The apparatus of claim 1 , wherein a support vector machine (SVM) is used for the machine training technique.

3. The apparatus of claim 1 , wherein the processor identifies whether the first surrounding vehicle detected by the front radar device and the second surrounding vehicle detected by the corner radar device are the same vehicle by using the first surrounding vehicle information and the second surrounding vehicle information.

4. The apparatus of claim 3 , wherein the processor fuses the first surrounding vehicle information and the second surrounding vehicle information when the first surrounding vehicle detected by the front radar device and the second surrounding vehicle detected by the corner radar device are the same vehicle.

5. The apparatus of claim 1 , wherein the processor decides that a current maneuver of the surrounding vehicle is a cut-in when a motion of the surrounding vehicle is dynamic, a previous maneuver of the surrounding vehicle is not a cut-in, and a decision result by the maneuver decision logic satisfies a reference condition.

6. The apparatus of claim 1 , wherein the processor decides that a current maneuver of the surrounding vehicle is driving when a motion of the surrounding vehicle is dynamic, a previous maneuver of the surrounding vehicle is not a cut-in, and a decision result by the maneuver decision logic fails to satisfy a reference condition.

7. The apparatus of claim 1 , wherein the processor decides that a current maneuver of the surrounding vehicle is a cut-in when a motion of the surrounding vehicle is stationary and a previous maneuver of the surrounding vehicle is a cut-in.

8. The apparatus of claim 1 , wherein the processor decides that a current maneuver of the surrounding vehicle is a stop when a motion of the surrounding vehicle is stationary and a previous maneuver of the surrounding vehicle is not a cut-in.

9. The apparatus of claim 1 , wherein the processor decides that a current maneuver of the surrounding vehicle is a cut-in when a motion of the surrounding vehicle is dynamic, a previous maneuver of the surrounding vehicle is a cut-in, and a distance between the vehicle and the surrounding vehicle is less than a reference value.

10. The apparatus of claim 1 , wherein the processor decides that a current maneuver of the surrounding vehicle is driving when a motion of the surrounding vehicle is dynamic, a previous maneuver of the surrounding vehicle is a cut-in, and a distance between the vehicle and the surrounding vehicle is more than a reference value.

11. A method for deciding a maneuver of a surrounding vehicle, the method comprising:
obtaining first surrounding vehicle information for a first surrounding vehicle relative to a vehicle by using a front radar device and second surrounding vehicle information for a second surrounding vehicle by using a corner radar device;
setting a gate based on a location of the first surrounding vehicle;
identifying whether the second surrounding vehicle is located within the set gate;
fusing the first surrounding vehicle information and the second surrounding vehicle information when the second surrounding vehicle is located within the set gate;
deciding a maneuver of the surrounding vehicle according to maneuver decision logic derived by a machine training technique by using the fused surrounding vehicle information;
deciding a motion of the surrounding vehicle by using the surrounding vehicle information; and
deciding a final maneuver of the surrounding vehicle based on a motion decision result of the surrounding vehicle and a maneuver decision result by the maneuver decision logic.

12. The method of claim 11 , wherein the obtaining of the surrounding vehicle information includes:
identifying whether the first surrounding vehicle detected by the front radar device and the second surrounding vehicle detected by the corner radar device are the same vehicle by using the first surrounding vehicle information and the second surrounding vehicle information.

13. The method of claim 12 , wherein the obtaining of the surrounding vehicle information includes:
generating the surrounding vehicle information by fusing the first surrounding vehicle information and the second surrounding vehicle information when the first surrounding vehicle detected by the front radar device and the second surrounding vehicle detected by the corner radar device are the same vehicle.

14. The method of claim 11 , wherein a support vector machine (SVM) is used for the machine training technique.

15. The method of claim 11 , wherein the deciding of the final maneuver of the surrounding vehicle includes:
deciding that a current maneuver of the surrounding vehicle is a cut-in when a motion of the surrounding vehicle is dynamic, a previous maneuver of the surrounding vehicle is not a cut-in, and a decision result by the maneuver decision logic satisfies a reference condition.

16. The method of claim 11 , wherein the deciding of the final maneuver of the surrounding vehicle includes:
deciding that a current maneuver of the surrounding vehicle is driving when a motion of the surrounding vehicle is dynamic, a previous maneuver of the surrounding vehicle is not a cut-in, and a decision result by the maneuver decision logic fails to satisfy a reference condition.

17. The method of claim 11 , wherein the deciding of the final maneuver of the surrounding vehicle includes:
deciding that a current maneuver of the surrounding vehicle is a cut-in when a motion of the surrounding vehicle is stationary and a previous maneuver of the surrounding vehicle is a cut-in.

18. The method of claim 11 , wherein the deciding of the final maneuver of the surrounding vehicle includes:
deciding that a current maneuver of the surrounding vehicle is a stop when a motion of the surrounding vehicle is stationary and a previous maneuver of the surrounding vehicle is not a cut-in.

19. The method of claim 11 , wherein the deciding of the final maneuver of the surrounding vehicle includes:
deciding that a current maneuver of the surrounding vehicle is a cut-in when a motion of the surrounding vehicle is dynamic, a previous maneuver of the surrounding vehicle is a cut-in, and a distance between the vehicle and the surrounding vehicle is less than a reference value.

20. The method of

Assignees

• Hyundai Motor Co Ltd
• Kia Motors Corp
• Univ Ajou Ind Academic Coop Found

Inventors

• Lee Hoon
• Jang Hyung Sun
• Yang Uk Il
• Won Ji Eun
• Han Seul Ki
• Yun Bo Young
• Song Bong Sob
• Lee Sung Woo
• Lee Sang Won

Classifications

• G01S2013/932

  using own vehicle data, e.g. ground speed, steering wheel direction · CPC title

• G06N20/00

  Machine learning · CPC title

• G01S13/726

  Multiple target tracking · CPC title

• B60W60/0027

  using trajectory prediction for other traffic participants · CPC title

• G01S2013/93274

  on the side of the vehicles · CPC title