Predicting yield behaviors

1. A method for merge scenario detection by an autonomous vehicle (AV), comprising:

identifying, based on map information, a merge scenario including a crossing lane and a merging lane;

identifying an interacting object-pair of the merge scenario including a crossing object and a merging object;

generating a crossing yield hypothesis for the crossing object, the crossing yield hypothesis indicating a first intention of the crossing object to yield a right of way to the merging object;

generating a crossing no-yield hypothesis for the crossing object, the crossing no-yield hypothesis indicating a second intention of the crossing object to not yield the right of way to the merging object;

determining a crossing yield reference driveline for the crossing yield hypothesis;

determining a crossing no-yield reference driveline for the crossing no-yield hypothesis;

providing at least one of a crossing yield trajectory corresponding to the crossing yield hypothesis or a crossing no-yield trajectory corresponding to the crossing no-yield hypothesis; and

determining a trajectory of the AV based on the at least one of the crossing yield trajectory or the crossing no-yield trajectory.

2. The method of claim 1 , wherein identifying the interacting object-pair of the merge scenario comprises:

identifying, based on first observation data received from sensors of the AV, the crossing object, wherein the crossing object is in the crossing lane; and

identifying the merging object, wherein the merging object is in the merging lane.

3. The method of claim 2 , wherein identifying the merging object comprises:

identifying, based on second observation data received from sensors of the AV, the merging object.

4. The method of claim 1 , wherein the AV is identified as the merging object.

5. The method of claim 4 , further comprising:

determining a crossing yield likelihood that the crossing object follows the crossing yield hypothesis and a crossing no-yield likelihood that the crossing object follows the crossing no-yield hypothesis.

6. The method of claim 5 , further comprising:

in response to the crossing yield likelihood being below a first threshold, determining at least one of a speed or an acceleration profile to operate the AV to merge onto the merging lane.

7. The method of claim 5 , further comprising:

in response to the crossing no-yield likelihood exceeding a second threshold, operating the AV to stop.

8. The method of claim 1 , wherein providing the at least one of the crossing yield trajectory corresponding to the crossing yield hypothesis or the crossing no-yield trajectory corresponding to the crossing no-yield hypothesis comprises:

predicting the crossing yield trajectory using a motion model of the crossing object, such that a current predicted path of the crossing object coincides with the crossing yield reference driveline.

9. The method of claim 8 , wherein the crossing yield trajectory is provided for a predetermined future length of time.

10. The method of claim 9 , wherein the predetermined future length of time is 6 seconds.

11. A system for merge scenario detection by an autonomous vehicle (AV), comprising:

a memory; and

a processor, the memory includes instructions executable by the processor to:

identify, based on map information, a merge scenario including a crossing lane and a merging lane;

identify an interacting object-pair of the merge scenario including a crossing object and a merging object;

generate a crossing yield hypothesis for the crossing object, the crossing yield hypothesis indicating a first intention of the crossing object to yield a right of way to the merging object;

generate a crossing no-yield hypothesis for the crossing object, the crossing no-yield hypothesis indicating a second intention of the crossing object to not yield the right of way to the merging object;

determine a crossing yield reference driveline for the crossing yield hypothesis;

determine a crossing no-yield reference driveline for the crossing no-yield hypothesis;

provide at least one of a crossing yield trajectory corresponding to the crossing yield hypothesis or a crossing no-yield trajectory corresponding to the crossing no-yield hypothesis; and

determine a trajectory of the AV based on the at least one of the crossing yield trajectory or the crossing no-yield trajectory.

12. The system of claim 11 , wherein to identify the interacting object-pair of the merge scenario comprises to:

identify, based on first observation data received from sensors of the AV, the crossing object, wherein the crossing object is in the crossing lane; and

identify the merging object, wherein the merging object is in the merging lane.

13. The system of claim 12 , wherein to identify the merging object comprises to:

identify, based on second observation data received from sensors of the AV, the merging object.

14. The system of claim 11 , wherein the AV is identified as the merging object.

15. The system of claim 14 , wherein the instructions further include instructions to:

determine a crossing yield likelihood that the crossing object follows the crossing yield hypothesis and a crossing no-yield likelihood that the crossing object follows the crossing no-yield hypothesis.

16. The system of claim 15 , wherein the instructions further include instructions to:

in response to the crossing yield likelihood being below a first threshold, determining at least one of a speed or an acceleration profile to operate the AV to merge onto the merging lane.

17. The system of claim 15 , wherein the instructions further include instructions to:

in response to the crossing no-yield likelihood exceeding a second threshold, operating the AV to stop.

18. The system of claim 11 , wherein to provide the at least one of the crossing yield trajectory corresponding to the crossing yield hypothesis or the crossing no-yield trajectory corresponding to the crossing no-yield hypothesis comprises to:

predict the crossing yield trajectory using a motion model of the crossing object, such that a current predicted path of the crossing object coincides with the crossing yield reference driveline.

19. The system of claim 18 , wherein the crossing yield trajectory is provided for a predetermined future length of time.

20. A system for merge scenario detection by an autonomous vehicle (AV), comprising:

a memory; and

a processor, the memory includes instructions executable by the processor to:

in response to detecting a merge scenario, identify an interacting object pair including a merging object and a crossing object, wherein the AV is the merging object;

generate a yield hypothesis and a no-yield hypothesis;

compute a yield reference path corresponding to the yield hypothesis and a no-yield reference path corresponding to the no-yield hypothesis;

determine a yield likelihood of the yield hypothesis and a no-yield likelihood of the no-yield hypothesis; and

operate the AV to merge or to wait until the merge scenario is no longer detected based on the yield likelihood and the no-yield likelihood.