Methods and systems for autonomous vehicle inference of routes for actors exhibiting unrecognized behavior

1. A method, comprising:

by one or more electronic devices of an autonomous vehicle:

identifying an actor in an environment of the autonomous vehicle that is exhibiting unrecognized behavior or that has exhibited unrecognized behavior within a certain time period,

generating a circle associated with the actor, wherein the circle has a radius that is a function of a velocity of the actor,

identifying one or more target points associated with the actor, wherein each target point represents a point along a possible route of the actor, wherein each target point is located along a circumference of the circle,

assigning a score to each target point,

selecting the target point associated with the lowest score, and

generating a reference path from the actor to the selected target point.

2. The method of claim 1 , further comprising:

using the reference path to predict a predicted path of the actor; and

adjusting one or more driving operations of the autonomous vehicle based on the predicted path of the actor that is predicted.

3. The method of claim 1 , wherein identifying an actor in an environment of the autonomous vehicle that is exhibiting unrecognized behavior or that has exhibited unrecognized behavior within a certain time period comprises:

comparing a behavior of the actor to one or more policies of recognized behavior for the autonomous vehicle; and

identifying the behavior as an unrecognized behavior if the behavior does not correspond to the one or more polices.

4. The method of claim 1 , wherein generating a circle associated with the actor comprises generating the circle having the radius, wherein the radius represents a distance at which it is predicted that the actor will merge back into the flow of traffic.

5. The method of claim 1 , wherein identifying one or target points associated with the actor comprises:

identifying one or more lanes located in front of the actor relative to its heading, wherein each lane comprises a centerline; and

for each identified lane, determining an intersection of the circle with the centerline of the lane.

6. The method of claim 1 , wherein assigning a score to each target point comprises

determining a heading discrepancy value;

determining a curvature parameter;

determining a lateral acceleration of the actor; and

summing the heading discrepancy value, the curvature parameter, and the lateral acceleration.

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

8. The method of claim 6 , wherein determining a curvature parameter comprises applying a function to a curvature associated with traversing a path between the actor and the target point.

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

10. The method of claim 6 , wherein determining a lateral acceleration of the actor comprises:

identifying a curvature of a path between the actor and the target point;

determining a velocity of the actor;

squaring the velocity of the actor to generate a squared value; and

determining a product of the curvature and the squared value.

11. The method of claim 1 , wherein generating a reference path from the actor to the selected target point comprises generating a spline between the actor and the target point.

12. A method, comprising:

by one or more electronic devices of an autonomous vehicle:

identifying an actor in an environment of the autonomous vehicle that is exhibiting unrecognized behavior or that has exhibited unrecognized behavior within a certain time period,

determining whether the actor was following a lane within a certain time period and is now not occupying any lane,

in response to determining that the actor was following the lane within the time period and is now not occupying any lane, predicting that the actor will move forward along it current path,

in response to not determining that the actor was following the lane within the time period and is now not occupying any lane:

generating a circle associated with the actor, wherein the circle has a radius that is a function of a velocity of the actor,

identifying one or more target points associated with the actor, wherein each target point represents a point along a possible route of the actor, wherein each target point is located along a circumference of the circle,

assigning a score to each target point,

selecting the target point associated with the lowest score, and

generating a reference path from the actor to the selected target point.

13. A system, comprising:

one or more electronic devices of an autonomous vehicle;

a computer-readable storage medium comprising one or more programming instructions that, when executed, cause the one or more electronic devices to:

identify an actor in an environment of the autonomous vehicle that is exhibiting unrecognized behavior or that has exhibited unrecognized behavior within a certain time period,

generate a circle associated with the actor, wherein the circle has a radius that is a function of a velocity of the actor,

identify one or more target points associated with the actor, wherein each target point represents a point along a possible route of the actor, wherein each target point is located along a circumference of the circle,

assign a score to each target point,

select the target point associated with the lowest score, and

generate a reference path from the actor to the selected target point.

14. The system of claim 13 , wherein the computer-readable storage medium further comprises one or more programming instruction that, when executed, cause the one or more electronic devices to:

use the reference path to predict a predicted path of the actor; and

adjust one or more driving operations of the autonomous vehicle based on the predicted path of the actor that is predicted.

15. The system of claim 13 , wherein the one or more programming instructions that, when executed, cause the one or more electronic devices to identify an actor in an environment of the autonomous vehicle that is exhibiting unrecognized behavior or that has exhibited unrecognized behavior within a certain time period comprise one or more programming instructions that, when executed, cause the one or more electronic devices to:

compare a behavior of the actor to one or more policies of recognized behavior for the autonomous vehicle; and

identify the behavior as an unrecognized behavior if the behavior does not correspond to the one or more polices.

16. The system of claim 13 , wherein the one or more programming instructions that, when executed, cause the one or more electronic devices to generate a circle associated with the actor comprise one or more programming instructions that, when executed, cause the one or more electronic devices to generate the circle having the radius, wherein the radius represents a distance at which it is predicted that the actor will merge back into the flow of traffic.

17. The system of claim 13 , wherein the one or more programming instructions that, when executed, cause the one or more electronic devices to identify one or target points associated with the actor comprise one or more programming instructions that, when executed, cause the one or more electronic devices to:

identify one or more lanes located in front of the actor relative to its heading, wherein each lane comprises a centerline; and

for each identified lane, determine an intersection of the circle with the centerline of the lane.

18. The system of claim 13 , wherein the one or more programming instructions that, when executed, cause the one or more electronic devices to assign a score to each target point comprise the one or more programming instructions that, when executed, cause the one or more electronic devices to:

determine a heading discrepancy value;

determine a curvature parameter;

determine a lateral acceleration of the actor; and

sum the heading discrepancy value, the curvature parameter, and the lateral acceleration.

19. The system of claim 18 , wherein the one or more programming instructions that, when executed, cause the one or more electronic devices to determine a heading discrepancy value comprise one or more programming instructions that, when executed, cause the one or more electronic devices to determine a cosine distance at the target point.

20. The system of claim 18 , wherein the one or more programming instructions that, when executed, cause the one or more electronic devices to determine a curvature parameter comprise one or more programming instructions that, when executed, cause the one or more electronic devices to apply a function to a curvature associated with traversing a path between the actor and the target point.

21. The system of claim 20 , wherein the result of the function is a first value if the curvature of the path is too tight to traverse, wherein the result is a second value if the curvature of the path is not too tight to traverse.

22. The system of claim 18 , wherein the one or more programming instructions that, when executed, cause the one or more electronic devices to determine a lateral acceleration of the actor comprise one or more programming instructions that, when executed, cause the one or more electronic devices to:

identify a curvature of a path between the actor and the target point;

determine a velocity of the actor;

square the velocity of the actor to generate a squared value; and

determine a product of the curvature and the squared value.

23. The system of claim 13 , wherein the one or more programming instructions that, when executed, cause the one or more electronic devices to generate a reference path from the actor to the selected target point comprise one or more programming instructions that, when executed, cause the one or more electronic devices to generate a spline between the actor and the target point.