Methods and systems for performing outlet inference by an autonomous vehicle to determine feasible paths through an intersection

Patent metadata

Field	Value
Publication number	US-11358598-B2
Application number	US-202017060817-A
Country	US
Kind code	B2
Filing date	Oct 1, 2020
Priority date	Oct 1, 2020
Publication date	Jun 14, 2022
Grant date	Jun 14, 2022

Abstract

An autonomous vehicle identifies an intersection, identifies an object in proximity to the intersection, identifies a plurality of outlets of the intersection, and, for each outlet, identifies a polyline associated with the outlet, identifies a target point along the polyline, and determines a constant curvature path from the object to the target point. The system determines a score associated with each outlet based at least in part on the constant curvature path of the outlet, generates a pruned set of outlets that includes one or more of the outlets from the plurality of outlets based on its score, and for each outlet in the pruned set, generates a reference path from the object to the target point of the outlet.

First claim

What is claimed is:

1. A method, comprising:
by one or more electronic devices of an autonomous vehicle:
identifying an intersection,
identifying an object in proximity to the intersection,
identifying a plurality of outlets of the intersection,
for each outlet:
identifying a polyline associated with the outlet,
identifying a target point along the polyline, and
determining a constant curvature path from the object to the target point,

determining a score associated with each outlet based at least in part on the constant curvature path of the outlet,
generating a pruned set of outlets that includes one or more of the outlets from the plurality of outlets based on its score, and
for each outlet in the pruned set, generating a reference path from the object to the target point of the outlet.

2. The method of claim 1 , further comprising:
using one or more of the generated reference paths to predict a trajectory of the object through the intersection; and
adjusting one or more driving operations of the autonomous vehicle based on the trajectory of the object through the intersection that is predicted.

3. The method of claim 1 , wherein identifying a target point along the polyline comprises identifying a target point located a certain distance from an end of the outlet.

4. The method of claim 1 , wherein determining a constant curvature path from the object to the target point comprises fitting a circle to the object and the target point in an object frame by:
locating a center of rotation of the object at an origin position;
applying a point constraint and a slope constraint on the circle such that the circle passes through the origin position; and
applying a second point constraint on the circle such that the circle passes through the target point.

5. The method of claim 1 , wherein determining a score associated with each outlet based at least in part on the constant curvature path of the outlet comprises:
determining a heading discrepancy value;
determining a curvature parameter;
determining a lateral acceleration of the object; and
summing the heading discrepancy value, the curvature parameter, and the lateral acceleration.

6. The method of claim 5 , wherein determining a heading discrepancy value comprises determining a cosine distance at the target point of the outlet.

7. The method of claim 5 , wherein determining a curvature parameter comprises applying a function to a curvature of the constant curvature path.

8. The method of claim 7 , wherein the result of the function is a first value if the curvature of the constant curvature path is too tight to traverse the outlet, wherein the result is a second value if the curvature of the constant curvature path is not too tight to traverse the outlet.

9. The method of claim 5 , wherein determining a lateral acceleration of the object comprises:
identifying a curvature of the constant curvature path of the outlet;
determining a velocity of the object;
squaring the velocity of the object to generate a squared value; and
determining a product of the curvature and the squared value.

10. The method of claim 1 , wherein generating a pruned set of outlets that includes one or more of the outlets from the plurality of outlets based on its score comprises identifying one or more outlets from the plurality of outlets whose score is less than a threshold value.

11. The method of claim 1 , wherein generating a reference path from the moving object to the target point of the outlet comprises generating a spline between the object and the target point in a map frame.

12. The method of claim 1 , wherein one or more of the reference paths are not aligned with mapped lane segments of a map frame.

13. The method of claim 1 , further comprising, for one or more of the generated reference paths:
determining one or more mapped lane segments that lead to the outlet associated with the reference path; and
determining whether one or more of the mapped lane segments match the reference path.

14. The method of claim 13 , further comprising:
in response to determining that only one of the mapped lane segments match the reference path, applying at least a portion of metadata associated with the one mapped lane segment to the reference path.

15. The method of claim 13 , further comprising:
in response to determining that multiple mapped lane segments match the reference path and that the multiple mapped lane segments have the same metadata, applying at least a portion of metadata associated with one of the multiple mapped lane segments to the reference path.

16. The method of claim 13 , further comprising:
in response to determining that there are no mapped lane segments that match the reference path, applying default metadata to the reference path, wherein the default metadata includes one or more instructions instructing the autonomous vehicle to yield to one or more actors in the intersection.

17. The method of claim 1 , wherein generating a reference path from the object to the target point of the outlet comprises generating a reference path that is different from one or more mapped lane segments through the intersection.

18. A system, comprising:
one or more electronic devices of an autonomous vehicle; and
a computer-readable storage medium comprising one or more programming instructions that, when executed, cause one or more of the electronic devices to:
identify an intersection,
identify an object in proximity to the intersection,
identify a plurality of outlets of the intersection,
for each outlet:
identify a polyline associated with the outlet,
identify a target point along the polyline, and
determine a constant curvature path from the object to the target point,

determine a score associated with each outlet based at least in part on the constant curvature path of the outlet,
generate a pruned set of outlets that includes one or more of the outlets from the plurality of outlets based on its score, and
for each outlet in the pruned set, generate a reference path from the object to the target point of the outlet.

19. The system of claim 18 , wherein the computer-readable storage medium further comprises one or more programming instructions that, when executed, cause one or more of the electronic devices to:
use one or more of the generated reference paths to predict a trajectory of the object through the intersection; and
adjust one or more driving operations of the autonomous vehicle based on the trajectory of the object through the intersection that is predicted.

20. The system of claim 18 , wherein the one or more programming instructions that, when executed, cause one or more of the electronic devices to identify a target point along the polyline comprise one or more programming instructions that, when executed, cause the one or more of the electronic devices to identify a target point located a certain distance from an end of the outlet.

21. The system of claim 18 , wherein the one or more programming instructions that, when executed, cause one or more of the electronic devices to determine a constant curvature path from the object to the target point comprise one or more programming instructions that, when executed, cause the one or more of the electronic devices to fit a circle to the object and the target point in an object frame by:
locating a center of rotation of the object at an origin position;
applying a point constraint and a slope constraint on the circle such that the circle pa

Assignees

• Argo Ai Llc

Inventors

• Hartnett Andrew T
• Carr G Peter K
• Foil Greydon
• Savtchenko Constantin

Classifications

• B60W2556/40

  High definition maps · CPC title

• B60W2554/4041

  Position · CPC title

• B60W60/0011

  involving control alternatives for a single driving scenario, e.g. planning several paths to avoid obstacles · CPC title

• B60W2554/402

  Type · CPC title

• B60W2554/4043

  Lateral speed · CPC title