Probabilistic object tracking and prediction framework

1. A method for world objects tracking and prediction by an autonomous vehicle (AV), comprising:

receiving, from sensors of the AV, a first observation data;

associating the first observation data with a first world object by:

determining, using a gating area around the first observation data, world objects within the gating area, the world objects being users of a vehicle transportation network and are maintained in a module of the AV; and

determining an association likelihood that the first observation data corresponds to the first world object;

determining hypotheses for the first world object, wherein a hypothesis corresponds to an intention of the first world object;

determining a respective hypothesis likelihood of each of the hypotheses indicating a likelihood that the first world object follows the intention;

determining, for at least one hypothesis of the hypotheses, a respective state, wherein the respective state comprises predicted positions of the first world object; and

in response to a query, providing a hypothesis of the hypotheses based on the respective hypothesis likelihood of each of the hypotheses.

2. The method of claim 1 , wherein a hypothesis of the hypotheses is determined using map information.

3. The method of claim 1 , further comprising:

receiving, from the sensors of the AV, a second observation data;

associating the second observation data with the first world object; and

updating the respective state based on the second observation data.

4. The method of claim 1 , further comprising:

predicting, based on a motion model of the first world object, a respective trajectory for at least some of the hypotheses.

5. The method of claim 4 , wherein predicting, based on the motion model of the first world object, the respective trajectory comprises:

determining a reference driveline for the first world object; and

predicting the respective trajectory, based on the motion model, such that a current predicted path of the first world object coincides with the reference driveline.

6. The method of claim 4 , further comprising:

receiving, from the sensors of the AV, a second observation data; and

associating the second observation data with one of the first world object and a second world object.

7. The method of claim 6 , further comprising:

updating the respective hypothesis likelihood of each of the hypotheses based on the association likelihood; and

updating the respective state of the first world object.

8. The method of claim 1 , further comprising:

receiving, from the sensors of the AV, second observation data;

on a condition that the second observation data meeting a criterion, associating the second observation data with a second world object; and

on a condition that the second observation data not meeting the criterion, ignoring the second observation data.

9. The method of claim 1 , wherein the association likelihood is determined in a case that the first world object is included in the world objects.

10. A system for world objects tracking and prediction by an autonomous vehicle (AV), comprising:

a processor configured to:

receive, from sensors of the AV, a first observation data;

associate the first observation data with a first world object by, wherein to associate the first observation data with the first world object comprises to:

determine, using a gating area around the first observation data, world objects within the gating area; and

in a case that the first world object is included in the world objects, determine an association likelihood that the first observation data corresponds to the first world object;

determine hypotheses for the first world object, wherein a hypothesis corresponds to a possible path of the first world object;

determine a respective hypothesis likelihood of each of the hypotheses indicating a likelihood that the first world object follows the possible path;

determine, for at least one hypothesis of the hypotheses, a respective state, wherein the respective state comprises predicted positions of the first world object; and

in response to a query, provide a hypothesis of the hypotheses based on the respective hypothesis likelihood of each of the hypotheses.

11. The system of claim 10 , wherein a hypothesis of the hypotheses is determined using map information.

12. The system of claim 10 , wherein the processor is further configured to:

receive, from the sensors of the AV, a second observation data;

associate the second observation data with the first world object; and

update the respective state based on the second observation data.

13. The system of claim 10 , wherein the processor is further configured to:

predict, based on a motion model of the first world object, a respective trajectory for at least some of the hypotheses.

14. The system of claim 13 , wherein to predict, based on the motion model of the first world object, the respective trajectory comprises to:

determine a reference driveline for the first world object; and

predict the respective trajectory, based on the motion model, such that a current predicted path of the first world object coincides with the reference driveline.

15. The system of claim 13 , wherein the processor is further configured to:

receive, from the sensors of the AV, a second observation data; and

associate the second observation data with one of the first world object and a second world object.

16. The system of claim 15 , wherein the processor is further configured to:

update the respective hypothesis likelihood of each of the hypotheses based on the association likelihood; and

update the respective state of the first world object.

17. The system of claim 10 , wherein the processor is further configured to:

receive, from the sensors of the AV, second observation data;

on a condition that the second observation data meeting a criterion, associate the second observation data with a second world object; and

on a condition that the second observation data not meeting the criterion, ignore the second observation data.

18. A non-transitory computer-readable storage medium, comprising executable instructions that, when executed by a processor, facilitate performance of instructions for world objects tracking and prediction by an autonomous vehicle (AV), the instructions comprising instructions comprising to:

create and maintain respective hypotheses for real-world objects, wherein the world objects being, along with the AV, users of a vehicle transportation network;

predict respective trajectories for the real-world object;

associate a received sensor observation with at least one of the real-world objects by instructions to:

determine, using a gating area around the received sensor observation, world objects within the gating area; and

in a case that the at least one of the real-world objects is included in the world objects, determine an association likelihood that the received sensor observation corresponds to the at least one of the real-world objects;

update a respective likelihood of at least some of the respective hypotheses; and

update a respective state associated with at least some of the respective hypotheses.

19. The non-transitory computer-readable storage medium of claim 18 , wherein the instructions further comprise instructions to:

receive a new sensor observation; and

create a new real-world object based on the new sensor observation.

20. The non-transitory computer-readable storage medium of claim 18 , wherein the instructions further comprise instructions to:

receive, from the sensors of the AV, second observation data;

when the second observation data meet a criterion, associate the second observation data with a second world object; and

when the second observation data do not meet the criterion, ignore the second observation data.