3D occlusion reasoning for accident avoidance

1. A method for use in traversing a vehicle transportation network by a first vehicle, the method comprising:

identifying an occlusion in the vehicle transportation network;

identifying, for a world object that is on a first side of the occlusion, a visibility grid on a second side of the occlusion,

wherein the second side of the occlusion is different from the first side;

determining, based on the visibility grid, whether other world objects exist on the second side of the occlusion; and

altering a driving behavior of the first vehicle responsive to determining that at least one world object exists on the second side of the occlusion.

2. The method of claim 1 , wherein the visibility grid is identified with respect to a region of interest that is a predefined distance above ground.

3. The method of claim 1 , wherein the first vehicle is the world object that is on the first side of the occlusion, and wherein altering the driving behavior of the first vehicle responsive to determining that the at least one world object exists on the second side of the occlusion comprises:

determining that the at least one world object is a second vehicle and that the second vehicle obstructs a first trajectory of the first vehicle; and

altering the driving behavior of the first vehicle so as to avoid a collision.

4. The method of claim 1 , wherein altering the driving behavior of the first vehicle responsive to determining that the at least one world object exists on the second side of the occlusion comprises:

determining that the first vehicle is in a blind spot of a second vehicle, wherein the first side corresponding to a visible area according to a gaze model of a driver of the second vehicle and the second side corresponding to a non-visible area according to the gaze model, and wherein the second vehicle is the world object; and

altering the driving behavior of the first vehicle so that the first vehicle is no longer in the blind spot of the second vehicle.

5. The method of claim 4 , wherein altering the driving behavior of the first vehicle so that the first vehicle is no longer in the blind spot of the second vehicle comprises:

controlling the first vehicle using at least one of a slow down action, a speed up action, or a lane change action.

6. The method of claim 4 , wherein determining that the first vehicle is in the blind spot of the second vehicle comprises:

determining at least a location or a pose of the second vehicle; and

determining a visibility model around the at least the location or the pose of the second vehicle based on a nominal field of view of a driver of the second vehicle.

7. The method of claim 1 , wherein the occlusion is identified using map information.

8. The method of claim 1 , wherein identifying, for the world object that is on the first side of the occlusion, the visibility grid on the second side of the occlusion comprises:

performing 3-dimensional ray-tracing through a 3-dimensional volume of the second side; and

marking a region of the second side as either observed or unobserved based on the 3-dimensional ray-tracing.

9. The method of claim 8 , performing the 3-dimensional ray-tracing through the 3-dimensional volume of the second side comprises:

removing ground points from sensor information.

10. A method for use in traversing a vehicle transportation network, the method comprising:

identifying a first trajectory of a first world object in the vehicle transportation network;

identifying a visibility grid of the first world object;

identifying, using the visibility grid, a second world object that is invisible to the first world object; and

in response to determining that the first world object is predicted to collide with the second world object, alerting at least one of the first world object or the second world object.

11. The method of claim 10 , wherein alerting at least one of the first world object or the second world object comprises:

emitting at least one of an audible or visual notification.

12. The method of claim 10 , wherein the first world object is identified using a component of the vehicle transportation network.

13. An apparatus for use in traversing a vehicle transportation network by a vehicle, the apparatus comprising:

a non-transitory computer readable medium; and

a processor configured to execute instructions stored on the non-transitory computer readable medium to:

identify an occlusion in the vehicle transportation network, wherein the vehicle is on a first side of the occlusion;

determine a visibility grid on a second side of the occlusion, wherein the second side is different from the first side; and

in response to determining, based on the visibility grid, that an external object exists within the second side, alter a driving behavior of the vehicle.

14. The apparatus of claim 13 , wherein the instructions further comprise instructions to:

in response to determining that the external object does not exist within the second side, not alter the driving behavior of the vehicle.

15. The apparatus of claim 14 ,

wherein the visibility grid comprises an unobserved region, and

wherein to determine that the external object does not exist within the second side comprises to:

determine that the external object does not exist based on a size of the unobserved region.

16. The apparatus of claim 15 , wherein to determine that the external object does not exist based on the size of the unobserved region comprises to:

determine that the external object does not exist based on changes to the size of the unobserved region over time.

17. The apparatus of claim 13 , wherein to alter the driving behavior of the vehicle comprises to:

in response to determining that a first trajectory of the external object obstructs the first trajectory of the vehicle, alter the driving behavior of the vehicle so as to avoid a collision.

18. The apparatus of claim 13 , wherein to determine the visibility grid comprises to:

perform 3-dimensional ray-tracing through a 3-dimensional volume of the second side; and

mark an area of the second side as non-visible based on the 3-dimensional ray-tracing.

19. The apparatus of claim 18 , wherein to perform the 3-dimensional ray-tracing through the 3-dimensional volume of the second side comprises to:

remove ground points from sensor information.

20. The apparatus of claim 13 , wherein the visibility grid includes a first region that is obscured and a second region that is unobscured.