System and method for lane vehicle localization with lane marking detection and likelihood scoring

• US 9,483,700 B1
• Filed: 05/13/2015
• Issued: 11/01/2016
• Est. Priority Date: 05/13/2015
• Status: Active Grant

• Current Assignee: Honda Motor Co., Ltd. (Honda Motor Company)
• Original Assignee: Honda Motor Co., Ltd. (Honda Motor Company)
• Inventors: Chen, Yuncong, Murarka, Aniket
• Primary Examiner(s): Fitzpatrick, Atiba O
• Application Number:
  US14/710,785
• Publication Number:
  US 20160335507A1
• Time in Patent Office: 538 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06V 20/588 Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road

Abstract

A computer-implemented method and a system for vehicle localization. The method includes identifying expected lane lines based on a lane level map and a position of a vehicle and identifying observed lane lines based on images captured by an imaging device of the vehicle and the position of the vehicle. The method includes matching the expected lane lines to the observed lane lines and determining a likelihood score based on the matching. The method includes determining a best vehicle pose based on the likelihood score.

Claims

1. A computer-implemented method for vehicle localization, comprising:

identifying expected lane lines based on a lane level map and a position of a vehicle;

identifying observed lane lines based on images captured by an imaging device of the vehicle and the position of the vehicle;

matching the expected lane lines to the observed lane lines, wherein matching the expected lane lines to the observed lane lines includes determining one or more matched line pairs, each pair including an expected lane line matched to an observed lane line, determining a number of matched observed lane lines, and determining a number of unmatched observed lane lines;

determining a likelihood score based on the matching, wherein determining the likelihood score is based on a difference between the expected lane line and the observed lane line in each matched pair and the number of unmatched observer lane lines; and

determining a best vehicle pose based on the likelihood score.

2. The computer-implemented method of claim 1, wherein identifying observed lane lines includes applying a lane detection filter to the images.

3. The computer-implemented method of claim 1, wherein the likelihood score is based on a difference between the expected lane line and the observed lane line in the matched line pair, wherein the difference is based on a distance between an end point of the expected lane line and an end point of the observed lane line.

4. The computer-implemented method of claim 1, wherein the likelihood score is based on a difference between the expected lane line and the observed lane line in the matched line pair, wherein the difference is based on an angle between the expected lane line and the observed lane line.

5. The computer-implemented method of claim 1, including retrieving one or more vehicle poses, and for each vehicle pose, determining the likelihood score for the vehicle pose and determining the best vehicle pose based on the likelihood score.

6. A system for vehicle localization, comprising:

an odometry device for determining a position of a vehicle;

a lane level map stored at a database, the lane level map based on the position of the vehicle;

an imaging device for capturing images in an area around the vehicle based on the position of the vehicle; and

a processor operatively coupled for computer communication to the odometry device, the database and the imaging device, the processor including,

a motion module that retrieves expected data from the lane level map, the expected data identifying expected lane lines based on the position of the vehicle and determines a pose of a vehicle;

an observation module that retrieves observed data from the images, the observed data identifying observed lane lines based on the position and the pose of the vehicle;

a matching module that matches the expected lane lines to the observed lane lines and determines a number of matched observed lines and a number of unmatched observed lines; and

a localization module that determines a likelihood score of the pose of the vehicle based on the matching, wherein the localization module determines the likelihood score based on the number of matched of observed lines and the number of unmatched observed lines, and the localization module determines a best vehicle pose based on the likelihood score of the pose.

7. The system of claim 6, wherein the matching module determines one or more matched line pairs, each pair including an expected lane line matched to an observed lane line.

8. The system of claim 7, wherein the localization module determines the likelihood score for each matched line pair based on a difference between the expected lane line and the observed lane line in the matched line pair.

9. The system of claim 8, wherein the localization module determines the difference as a distance between an end point of the expected lane line and an end point of the observed lane line.

10. A non-transitory computer readable medium comprising instructions that when executed by a processor perform a method for vehicle localization, comprising:

identifying expected lane lines based on a lane level map and a position of a vehicle;

identifying observed lane lines based on images captured by an imaging device of the vehicle and the position of the vehicle;

matching the expected lane lines to the observed lane lines, wherein matching the expected lane lines to the observed lane lines includes determining one or more matched line pairs, each pair including an expected lane line matched to an observed lane line, and

determining a number of unmatched observed lane lines;

determining a likelihood score for each of the one or more matched line pairs based on a difference between the expected lane line and the observed lane line in each matched pair and the number of unmatched observer lane lines; and

determining a best vehicle pose based on the likelihood score.

11. The non-transitory computer readable medium of claim 10, wherein the difference is based on a distance between an end point of the expected lane line and an end point of the observed lane line.

12. The non-transitory computer readable medium of claim 11, including retrieving one or more vehicle poses, and for each vehicle pose, determining the likelihood score for the vehicle pose and determining the best vehicle pose based on the likelihood score.