Crowdsourcing a sparse map for autonomous vehicle navigation

Patent metadata

Field	Value
Publication number	US-12147242-B2
Application number	US-202117349124-A
Country	US
Kind code	B2
Filing date	Jun 16, 2021
Priority date	Jul 21, 2016
Publication date	Nov 19, 2024
Grant date	Nov 19, 2024

Abstract

Systems and methods are provided for crowdsourcing a sparse map for autonomous vehicle navigation. In one implementation, a non-transitory computer-readable medium may include a sparse map for autonomous vehicle navigation along a road segment. The sparse map may include at least one line representation of a road surface feature extending along the road segment, each line representation representing a path along the road segment substantially corresponding with the road surface feature, and wherein the road surface feature is identified through image analysis of a plurality of images acquired as one or more vehicles traverse the road segment and a plurality of landmarks associated with the road segment.

First claim

What is claimed is:

1. A non-transitory computer-readable medium including a sparse map for navigating along a road segment, the sparse map comprising:
at least one line representation of a road surface feature extending along the road segment, each line representation substantially corresponding with the road surface feature, the at least one line representation being determined by:
receiving first data generated by performing a first image analysis on a plurality of images, the plurality of images being acquired as a plurality of vehicles traverse the road segment; and
aligning the first data;

a target trajectory representing an intended path for a host vehicle when traveling along the road segment, the target trajectory being determined by:
determining, based on the first data, a plurality of trajectories traveled by the plurality of vehicles along the road segment; and
clustering the plurality of trajectories; and

a plurality of landmarks associated with the road segment, wherein the landmarks are determined by:
receiving second data generated by performing a second image analysis on the plurality of images; and
aligning the second data.

2. The non-transitory computer-readable medium of claim 1 , wherein the sparse map further comprises at least one of: a road width profile, a road roughness profile, a traffic line spacing profile, or a road condition.

3. The non-transitory computer-readable medium of claim 1 , wherein aligning the first or second data comprises averaging together multiple position measurements associated with a particular feature.

4. The non-transitory computer-readable medium of claim 3 , wherein aligning the first data comprises averaging a cluster of trajectories by:
determining a reference frame trajectory; and
determining a transformation that maps the cluster of trajectories to the reference frame trajectory.

5. The non-transitory computer-readable medium of claim 1 , wherein the target trajectory is based on two reconstructed trajectories of traversals of the plurality of vehicles along the road segment.

6. The non-transitory computer-readable medium of claim 1 , wherein:
the target trajectory is associated with a first lane of the road segment; and
the sparse map comprises another trajectory associated with a second lane of the road segment.

7. The non-transitory computer-readable medium of claim 1 , wherein the plurality of landmarks enable a host vehicle navigating according to the sparse map to determine a position of the host vehicle.

8. The non-transitory computer-readable medium of claim 1 , wherein the road surface feature includes at least one of a road edge or a lane marking.

9. The non-transitory computer-readable medium of claim 1 , wherein at least one of the plurality of landmarks includes a road sign.

10. The non-transitory computer-readable medium of claim 1 , wherein the at least one line representation comprises at least one polynomial representation.

11. The non-transitory computer-readable medium of claim 10 , wherein the at least one polynomial representation is a three-dimensional polynomial representation.

12. The non-transitory computer-readable medium of claim 1 , wherein the second image analysis comprises further determining the landmarks when a ratio of images in which the landmark appears to images in which the landmark does not appear exceeds a threshold.

13. The non-transitory computer-readable medium of claim 1 , wherein the first image analysis comprises identifying one or more edges in the plurality of images.

14. The non-transitory computer-readable medium of claim 1 , wherein the plurality of vehicles traversing the road segment are associated with a common heading or a lane assignment.

15. The non-transitory computer-readable medium of claim 1 , wherein the target trajectory is based on at least one environmental condition.

16. The non-transitory computer-readable medium of claim 1 , wherein the target trajectory is based on ego motion estimations of the plurality of vehicles.

17. The non-transitory computer-readable medium of claim 16 , wherein the ego motion estimations are based on rotation and translation estimation, the rotation and translation estimation being determined based on the first image analysis of the plurality of images.

18. The non-transitory computer-readable medium of claim 16 , wherein the ego motion estimations are based on an optical flow analysis of the plurality of images.

19. A method for generating a sparse map for navigating along a road segment, comprising:
determining at least one line representation of a road surface feature extending along the road segment by:
receiving first data generated by performing a first image analysis on a plurality of images, the plurality of images being acquired as a plurality of vehicles traverse the road segment; and
aligning the first data;

determining a target trajectory representing an intended path for a host vehicle when traveling along the road segment by:
determining, based on the first data, a plurality of trajectories traveled by the plurality of vehicles along the road segment; and
clustering the plurality of trajectories; and

determining a plurality of landmarks associated with the road segment by:
receiving second data generated by performing a second image analysis on the plurality of images; and
aligning the second data.

20. A system for generating a sparse map for navigating along a road segment, the system comprising:
at least one processor comprising circuitry and a memory, wherein the memory includes instructions that when executed by the circuitry cause the at least one processor to:
determine at least one line representation of a road surface feature extending along the road segment by:
receiving first data generated by performing a first image analysis on a plurality of images, the plurality of images being acquired as a plurality of vehicles traverse the road segment; and
aligning the first data;

determine a target trajectory representing an intended path for a host vehicle when traveling along the road segment by:
determining, based on the first data, a plurality of trajectories traveled by the plurality of vehicles along the road segment; and
clustering the plurality of trajectories; and

determine a plurality of landmarks associated with the road segment by:
receiving second data generated by performing a second image analysis on the plurality of images; and
aligning the second data.

Assignees

• Mobileye Vision Technologies Ltd

Inventors

• Fridman Ofer

Classifications

• G06T2207/30256

  Lane; Road marking · CPC title

• G06T2207/10032

  Satellite or aerial image; Remote sensing · CPC title

• G06T7/37

  using transform domain methods · CPC title

• G06T7/337

  involving reference images or patches · CPC title

• G06T7/11

  Region-based segmentation · CPC title