Methods and systems for improving maps

1. A method for improving map completeness, comprising:

obtaining sensor data from a plurality of sensors over time, the plurality of sensors covering a sensor field;

generating a map by fusing the obtained sensor data, wherein the generated map comprises a plurality of grid cells at least partially covered by the sensor field;

determining, using the obtained sensor data, completeness of at least one portion of the generated map by determining object completeness and coverage completeness of the map from the obtained sensor data; and

updating completeness of at least the portion of the generated map by reducing an amount of unknown areas of the at least one portion of the generated map using the determined object completeness and the determined coverage completeness,

wherein determining coverage completeness comprises

determining one or more unknown grid cells of the generated map, wherein the one or more unknown grid cells are grid cells for which there is insufficient sensor information.

2. The method of claim 1 , wherein determining object completeness comprises:

obtaining sensor data from ingress and egress sensors;

determining, from data obtained from ingress and egress sensors, objects entering and leaving at least one portion of the generated map;

determining, from the obtained sensor data, a current quantity of objects in the at least one portion of the generated map; and

tracking, from the obtained sensor data, each of one or more objects in the at least one portion of the generated map.

3. The method of claim 2 , wherein determining object completeness comprises determining disappearance of one or more objects from the at least portion of the generated map based on the tracking, the determined amount of objects and a net change of objects in the at least one portion of the generated map.

4. The method of claim 3 , wherein in response to determining one or more objects have disappeared from the at least one portion of the generated map, the method further comprises

determining a prediction space, wherein the prediction space comprises a set of grid cells for which the one or more determined disappeared objects can currently occupy.

5. The method of claim 4 , wherein determining the prediction space further comprises:

for each determined disappeared object, determining possible positions the disappearing object can currently occupy using past sensor data of each of the disappeared objects.

6. The method of claim 5 , wherein updating completeness of at least the portion of the generated map comprises:

eliminating unknown grid cells that are members of the determined prediction space.

7. The method of claim 6 , wherein eliminating unknown grid cells comprises assigning a known status to the eliminated unknown grid cells.

8. The method of claim 5 , wherein determining coverage completeness further comprises determining one or more known grid cells of the generated map, wherein a grid cell is known in response to determining from the obtained sensor data that the grid cell is occupied or unoccupied by an object.

9. The method of claim 1 , wherein one or more grid cells are determined to be unknown in response to determining the one or more grid cells are not covered by the sensor field.

10. The method of claim 1 , wherein one or more grid cells are determined to be unknown in response to determining the one or more grid cells are not covered due to failure of one or more of the plurality of sensors.

11. The method of claim 1 , wherein one or more grid cells are determined to be unknown in response to determining the one or more grid cells are not covered due to occlusion.

12. The method of claim 2 , wherein the object is a vehicle.

13. The method of claim 1 , further comprising obtaining an area of interest, wherein the at least one portion of the generated map is the obtained area of interest.

14. The method of claim 1 , further comprising, determining one or more completeness metrics of the at least one portion of the generated after improving the completeness.

15. The method of claim 2 , wherein determining object completeness comprises determining no objects are missing from the at least portion of the generated map based on the tracking, the determined amount of objects and a net change of objects in the at least one portion of the generated map.

16. The method of claim 15 , wherein in response to determining no objects are missing, determining all grid cells of the least one portion are known.

17. One or more computing devices comprising one or more processors and at least one non-transitory computer-readable storage medium including instructions that, when executed by the one or more processors, cause the one or more processors to:

obtain sensor data from a plurality of sensors over time, the plurality of sensors covering a sensor field;

generate a map by fusing the obtained sensor data, wherein the generated map comprises a plurality of grid cells at least partially covered by the sensor field;

determine, using the obtained sensor data, completeness of at least one portion of the generated map by determining object completeness and coverage completeness of the map from the obtained sensor data; and

update completeness of at least the portion of the generated map by reducing an amount of unknown areas of the at least one portion of the generated map using the determined object completeness and the determined coverage completeness,

wherein to determine coverage completeness comprises to

determine one or more unknown grid cells of the generated map, wherein the one or more unknown grid cells are grid cells for which there is insufficient sensor information.

18. The one or more computing devices of claim 17 , wherein the executed instructions cause the one or more processors to determine to object completeness by causing the one or more processors to:

obtain sensor data from ingress and egress sensors;

determine, from data obtained from ingress and egress sensors, objects entering and leaving at least one portion of the generated map;

determine, from the obtained sensor data, a current quantity of objects in the at least one portion of the generated map; and

track, from the obtained sensor data, each of one or more objects in the at least one portion of the generated map.

19. The one or more computing devices of claim 18 , wherein the executed instructions cause the one or more processors to determine to object completeness by causing the one or more processors to determine disappearance of one or more objects from the at least portion of the generated map based on the tracked objects, the determined amount of objects, and a net change of objects in the at least one portion of the generated map.