Method and system for operating an autonomous agent with incomplete environmental information

1. A method for operating an autonomous vehicle, the method comprising:

determining an environmental representation characterizing surroundings of the autonomous vehicle;

determining a set of unknown regions in the environmental representation;

selecting a set of virtual objects based on the set of unknown regions;

inserting the set of virtual objects into the set of unknown regions in the environmental representation to produce a modified environmental representation; and

operating the autonomous vehicle based on the modified environmental representation.

2. The method of claim 1 , wherein selecting the set of virtual objects based on the set of unknown regions comprises referencing a library of predetermined virtual objects based at least on a set of features associated with the set of unknown regions.

3. The method of claim 2 , wherein the set of features comprises at least one of:

a set of dimensions associated with the set of unknown regions; or

a characterized set of scenarios associated with the set of unknown regions.

4. The method of claim 3 , wherein at least one selected virtual object of the set has a largest virtual dimension permitted by a dimension of the set of dimensions associated with a corresponding unknown region of the set of unknown regions.

5. The method of claim 3 , wherein the characterized set of scenarios comprises at least one of a crosswalk and a bicycle lane.

6. The method of claim 5 , wherein in an event that an unknown region of the set is associated with a crosswalk, a corresponding selected virtual object inserted into the unknown region is configured with characteristics of a pedestrian.

7. The method of claim 5 , wherein in an event that unknown region of the set is associated with a bicycle lane, a corresponding selected virtual object inserted into the unknown region is configured with characteristics of a bicycle.

8. The method of claim 2 , wherein the selected set of virtual objects is further selected based on a second set of features associated with at least one of: the autonomous vehicle or a set of known objects present in the environmental representation.

9. The method of claim 1 , wherein each of the set of unknown regions is determined, at least in part, based on a location associated with each of the set of unknown regions relative to a predetermined set of regions in a labeled map.

10. The method of claim 9 , wherein the predetermined set of regions comprises a set of potential conflict zones, wherein each of the set of unknown regions overlaps with a potential conflict zone of the set of potential conflict zones.

11. The method of claim 1 , wherein the selected set of virtual objects is further determined based on a set of features associated with at least one of: the autonomous vehicle or a set of known objects present in the environmental representation.

12. The method of claim 11 , wherein the second set of features comprises a set of speeds.

13. The method of claim 1 , wherein selecting the set of virtual objects further comprises assigning a speed to each of the set of virtual objects.

14. The method of claim 13 , wherein a first speed assigned to a first virtual object of the set of virtual objects is selected to have a highest value that is above a probability threshold.

15. The method of claim 14 , wherein the probability threshold is determined based on at least one of: a speed of the autonomous vehicle, a set of speeds associated with a set of known objects in the environmental representation, or a set of traffic rules.

16. The method of claim 1 , wherein inserting the set of virtual objects into the set of unknown regions comprises determining a placement of each of the set of virtual objects into a corresponding unknown region of the set of unknown regions.

17. The method of claim 16 , wherein the placement is determined based on at least one of: a location of the autonomous vehicle and a lane line within the corresponding unknown region.

18. The method of claim 16 , wherein the placement is configured to minimize a distance of a virtual object of the set relative to the autonomous vehicle.

19. The method of claim 1 , further comprising refraining from inserting a virtual object of the set of virtual objects into an unknown region of the set of unknown regions based on historical information associated with a set of known objects in the environmental representation.

20. A system for operating an autonomous vehicle, the system comprising:

a set of sensors onboard the autonomous vehicle, the set of sensors configured to collect sensor data;

a processing subsystem in communication with the set of sensors, the processing subsystem configured to:

determine an environmental representation characterizing surroundings of the autonomous vehicle based on the sensor data;

determine a set of unknown regions in the environmental representation;

select a set of virtual objects based on the set of unknown regions;

insert the set of virtual objects into the set of unknown regions in the environmental representation to produce a modified environmental representation; and

operate the autonomous vehicle based on the modified environmental representation.