Semantic map with live updates

Patent metadata

Field	Value
Publication number	US-11226207-B2
Application number	US-201916718879-A
Country	US
Kind code	B2
Filing date	Dec 18, 2019
Priority date	Dec 18, 2019
Publication date	Jan 18, 2022
Grant date	Jan 18, 2022

Abstract

Systems, methods, and devices are disclosed for synchronizing map updates for a navigating autonomous vehicle. A management service on a managing node can receive an update to a portion of a first map, where a limited visibility map, which is a subset of the first map, can be determined based on a geographical area around an autonomous vehicle navigating a route. The limited visibility map can be synchronized among all the nodes while the autonomous vehicle is using the limited visibility map by the management service delivering, to a plurality of downstream nodes, the update to the limited visibility map when the update does not modify a portion of the limited visibility map, and by bypassing, to the plurality of downstream nodes, the update to the limited visibility map when the update modifies a portion of the limited visibility map.

First claim

What is claimed is:

1. A method comprising:
receiving, at a management service on a managing node, an update to a map used by an autonomous vehicle to navigate a route;
determining a partial map comprising a subset of the map used by the autonomous vehicle, wherein the partial map is defined by a geographical area around the autonomous vehicle navigating the route;
determining whether the update to the map includes a modification to the partial map that creates an inconsistency between a first version of the partial map prior to the update and a second version of the partial map after the update; and
synchronizing the partial map among a set of nodes of the autonomous vehicle while the autonomous vehicle is using the partial map to navigate the route, wherein synchronizing the partial map comprises delivering, to one or more downstream nodes of the set of nodes while the autonomous vehicle is in the geographical area associated with the partial map, the update to the map based on a determination that the update does not include the modification to the partial map.

2. The method of claim 1 , wherein the managing node is a router and the method further comprises:
queuing, at the router, a plurality of updates, wherein each update is associated with a respective version of the map;
determining a current location of the autonomous vehicle and a future location of the autonomous vehicle; and
delivering, to the one or more downstream nodes, a specific update associated with a specific version of the map based on a determination to publish the specific update;
wherein the determination to publish is based on the specific update modifying a portion of the map outside of a range of the current location and the future location of the autonomous vehicle.

3. The method of claim 1 , wherein the geographical area that determines the partial map is defined by:
determining a subset of raw first map data that represents a fixed area around a position of the autonomous vehicle within the map; and
extending the subset to include raw first map data outside of the fixed area associated with a semantic feature included within the fixed area;
wherein the semantic feature comprises one or more of lanes, lane boundaries, or drivable and undrivable areas.

4. The method of claim 3 , wherein extending the subset to include the semantic feature within the fixed area is further extended to include an additional semantic feature entirely outside the fixed area, wherein the additional semantic feature is connected to the semantic feature based on a relationship between the sematic feature and the additional semantic feature.

5. The method of claim 1 , wherein the partial map comprises a plurality of semantic features, each semantic feature represented in a form of a graph, and wherein the partial map is comprised of multiple layers of graphs representing each semantic feature.

6. The method of claim 1 , further comprising:
determining a second partial map comprising a second subset of the map;
determining whether one or more updates to the map include one or more modifications to the second partial map; and
synchronizing the second partial map among the set of nodes of the autonomous vehicle while the autonomous vehicle is using the second partial map to navigate the route, wherein synchronizing the second partial map comprises bypassing, to the one or more downstream nodes, the one or more updates to the map based on a determination that the one or more updates include the one or more modifications to the second partial map.

7. The method of claim 1 , wherein the downstream nodes are independent processes associated with sensors and a navigation service, and wherein the independent processes communicate with each other while the autonomous vehicle navigates the route, the method further comprising:
sending the update to a sandboxed cache when the update modifies a portion of the partial map such that the update can be applied at a subsequent time.

8. A non-transitory computer readable medium comprising instructions, the instructions, when executed by a computing system, cause the computing system to:
receive, at a management service on a managing node, an update to a map used by an autonomous vehicle to navigate a route;
determine a partial map comprising a subset of the map used by the autonomous vehicle, wherein the partial map is defined by a geographical area around the autonomous vehicle navigating the route;
determine whether the update to the map includes a modification to the partial map that creates an inconsistency between a first version of the partial map prior to the update and a second version of the partial map after the update; and
synchronize the partial map among a set of nodes of the autonomous vehicle while the autonomous vehicle is using the partial map to navigate the route, wherein synchronizing the partial map comprises delivering, to one or more downstream nodes of the set of nodes while the autonomous vehicle is in the geographical area associated with the partial map, the update to the map based on a determination that the update does not include the modification to the partial map.

9. The non-transitory computer readable medium of claim 8 , wherein the managing node is a router, and wherein the instructions further cause the computing system to:
queue, at the router, a plurality of updates, wherein each update is associated with a respective version of the map;
determine a current location of the autonomous vehicle and a future location of the autonomous vehicle; and
deliver, to the one or more downstream nodes, a specific update associated with a specific version of the map based on a determination to publish the specific update;
wherein the determination to publish is based on the specific update modifying a portion of the first map outside of a range of the current location and the future location of the autonomous vehicle.

10. The non-transitory computer readable medium of claim 8 , wherein the instructions that cause the computing system to determine a partial map further causes the computing system to:
determine a subset of raw first map data that represents a fixed area around a position of the autonomous vehicle within the first map; and
extend the subset to include raw first map data outside of the fixed area associated with a semantic feature included within the fixed area;
wherein the semantic feature comprises one or more of lanes, lane boundaries, or drivable and undrivable areas.

11. The non-transitory computer readable medium of claim 10 , wherein the instructions further cause the computing system to further extend the subset to include an additional semantic feature entirely outside the fixed area, wherein the additional semantic feature is connected to the semantic feature based on a relationship between the sematic feature and the additional semantic feature.

12. The non-transitory computer readable medium of claim 8 , wherein the partial map comprises a plurality of semantic features, each semantic feature represented in a form of a graph, and wherein the partial map is comprised of multiple layers of graphs representing each semantic feature.

13. The non-transitory computer readable medium of claim 8 , wherein the instructions further cause the computing system to:
determine a second partial map comprising a second subset of the map;
determine whether one or more updates to the map include one or more modifications to the second partial map; and
synchronize the second partial map among the set of nodes of the autonomous vehicle while the autonomous vehicle is using the second partial map to navigate the rou

Assignees

• Gm Cruise Holdings Llc

Inventors

• Sergeev Roman
• Robinson Andrew
• Marchioro Rech Lucio Rech

Classifications

• G05D1/246

  using environment maps, e.g. simultaneous localisation and mapping [SLAM] · CPC title

• G06F16/29

  Geographical information databases · CPC title

• G01C21/20

  Instruments for performing navigational calculations (G01C21/24, G01C21/26 take precedence) · CPC title

• G06V20/588

  Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road · CPC title

• G01C21/3859 Primary

  Differential updating map data · CPC title