Autonomous vehicle control using submaps
A system to use submaps to control operation of a vehicle is disclosed. A storage system may be provided with a vehicle to store a collection of submaps that represent a geographic area where the vehicle may be driven. A programmatic interface may be provided to receive submaps and submap updates independently of other submaps.
1. A computer-implemented method comprising:
retrieving, by one or more computing devices physically located onboard an autonomous vehicle, one or more submaps of a plurality of submaps representing an area of a road network for a geographic region;
detecting, by the one or more computing devices, a submap change condition, wherein the submap change condition comprises one or more of an environmental condition, an event occurrence, or a submap update;
retrieving, by the one or more computing devices, based at least in part on the submap change condition, one or more alternative submaps of the plurality of submaps, wherein the one or more alternative submaps represent the same area of the road network for the geographic region as the one or more submaps;
determining, by the one or more computing devices and based at least in part on a comparison of contemporaneous sensor data of the autonomous vehicle with corresponding sensor data previously stored in association with the one or more alternative submaps, a location and pose of the autonomous vehicle within the area of the road network for the geographic region; and
controlling, by the one or more computing devices and based at least in part on the location and pose of the autonomous vehicle, one or more operations of the autonomous vehicle.
2. The computer-implemented method of claim 1 , wherein determining the location and pose of the autonomous vehicle comprises comparing one or more objects or features of a scene of the area as represented by the contemporaneous sensor data with one or more objects or features of a previous scene of the area as represented by the corresponding sensor data to determine one or more of a spatial or geometric differential between the scene of the area as represented by the contemporaneous sensor data and the previous scene of the area as represented by the corresponding sensor data.
3. The computer-implemented method of claim 1 , wherein:
the contemporaneous sensor data comprises data, captured by one or more image sensors of the autonomous vehicle, representing a scene of the area;
the corresponding sensor data comprises data, captured by one or more corresponding image sensors, representing a previous scene of the area; and
determining the location and pose of the autonomous vehicle comprises comparing the data representing the scene with the data representing the previous scene.
4. The computer-implemented method of claim 1 , wherein:
the contemporaneous sensor data comprises data, captured by one or more light detection and ranging (LIDAR) sensors of the autonomous vehicle, representing a scene of the area;
the corresponding sensor data comprises data, captured by one or more corresponding LIDAR sensors, representing a previous scene of the area; and
determining the location and pose of the autonomous vehicle comprises comparing the data representing the scene with the data representing the previous scene.
5. The computer-implemented method of claim 1 , wherein:
the contemporaneous sensor data comprises data representing a point cloud representing a scene of the area;
the corresponding sensor data comprises data representing a point cloud representing a previous scene of the area; and
determining the location and pose of the autonomous vehicle comprises comparing the data representing the point cloud representing the scene with the data representing the point cloud representing the previous scene.
6. The computer-implemented method of claim 5 , wherein the data representing the point cloud representing the scene represents a point cloud spanning radially from a reference location in front of the autonomous vehicle.
7. The computer-implemented method of claim 5 , wherein the data representing the point cloud representing the scene represents a point cloud corresponding to one or more spaces alongside of the autonomous vehicle.
8. The computer-implemented method of claim 5 , wherein the data representing the point cloud representing the scene represents a point cloud corresponding to one or more spaces behind the autonomous vehicle.
9. The computer-implemented method of claim 5 , wherein the method comprises:
identifying, by the one or more computing devices, a plurality of different point clouds representing the previous scene; and
selecting, by the one or more computing devices and from amongst the plurality of different point clouds, the point cloud representing the previous scene.
10. The computer-implemented method of claim 9 , wherein selecting the point cloud representing the previous scene comprises selecting the point cloud representing the previous scene based at least in part on one or more of:
a lighting condition associated with the previous scene;
a weather condition associated with the previous scene;
a time of day associated with the previous scene; or
a season associated with the previous scene.
11. The computer-implemented method of claim 1 , wherein determining the location and pose of the autonomous vehicle comprises determining one or more of:
a lane the autonomous vehicle is using;
a distance of the autonomous vehicle from an edge of a road of the road network;
a distance of the autonomous vehicle from an edge of a lane of a road of the road network; or
a distance of travel from a point of reference for the one or more alternative submaps.
12. The computer-implemented method of claim 1 , wherein determining the location and pose of the autonomous vehicle comprises determining a location coordinate of the autonomous vehicle with respect to a particular submap of the one or more alternative submaps.
13. The computer-implemented method of claim 1 , wherein determining the location and pose of the autonomous vehicle comprises determining an orientation of the autonomous vehicle with respect to a particular road segment of the road network.
14. The computer-implemented method of claim 1 , wherein determining the location and pose of the autonomous vehicle comprises determining one or more of the location or pose of the autonomous vehicle based at least in part on one or more of a priority or weight indicating one or more of reliability or effectiveness of the corresponding sensor data for purposes of determining the one or more of the location or pose of the autonomous vehicle.
15. The computer-implemented method of claim 1 , wherein determining the location and pose of the autonomous vehicle comprises comparing the contemporaneous sensor data with the corresponding sensor data to determine one or more objects or features of the area that can form a basis for one or more of geometric or spatial comparison.
16. The computer-implemented method of claim 1 , wherein determining the location and pose of the autonomous vehicle comprises determining one or more of the location or pose of the autonomous vehicle with respect to a location at which the corresponding sensor data was captured.
17. An autonomous vehicle comprising:
one or more processors; and
a memory storing instructions that when executed by the one or more processors cause the autonomous vehicle to perform operations comprising:
retrieving one or more submaps of a plurality of submaps representing an area of a road network for a geographic region;
detecting a submap change condition, wherein the submap change condition comprises one or more of an environmental condition, and event occurrence, or a submap update;
retrieving, based at least in part on the submap change condition, one or more alternative submaps of the plurality of submaps referencing data representing a point cloud representing a previous scene of the same area of the road network for the geographic region represented by the one or more submaps;
receiving sensor data representing a point cloud representing a contemporaneous scene of the area of the road network; and
determining, based at least in part on a comparison of the data representing the point cloud representing the previous scene with the data representing the point cloud representing the contemporaneous scene, one or more of a location or pose of the autonomous vehicle.
18. The autonomous vehicle of claim 17 , wherein determining the one or more of the location or pose of the autonomous vehicle comprises comparing one or more objects or features of the previous scene as represented by the one or more alternative submaps with one or more objects or features of the contemporaneous scene as represented by the sensor data to determine one or more of a spatial or geometric differential between the previous scene as represented by the one or more alternative submaps and the contemporaneous scene as represented by the sensor data.
19. The autonomous vehicle of claim 17 , wherein:
the sensor data representing the point cloud representing the contemporaneous scene comprises data generated based at least in part on data captured by one or more image sensors of the autonomous vehicle; and
the data representing the point cloud representing the previous scene comprises data generated based at least in part on data captured by one or more corresponding image sensors.
20. The autonomous vehicle of claim 17 , wherein:
the sensor data representing the point cloud representing the contemporaneous scene comprises data generated based at least in part on data captured by one or more light detection and ranging (LIDAR) sensors of the autonomous vehicle; and
the data representing the point cloud representing the previous scene comprises data generated based at least in part on data captured by one or more corresponding LIDAR sensors.
21. One or more non-transitory computer-readable media comprising instructions that when executed by one or more computing devices physically located onboard an autonomous vehicle cause the one or more computing devices to perform operations comprising:
retrieving one or more submaps of a plurality of submaps representing an area of a road network for a geogrpahic region;
detecting a submap change condition, wherein the submap change condition comprises one or more of an environmental condition, an event occurrence, or a submap update;
retrieving, based at least in part on the submap change condition, one or more alternative submaps of the plurality of submaps referencing data representing a previous scene of the same area of the road network for the geographic region as the one or more submaps;
receiving sensor data representing a contemporaneous scene of the area; and
determining one or more of a location or pose of the autonomous vehicle by comparing one or more objects or features of the previous scene as represented by the one or more alternative submaps with one or more objects or features of the contemporaneous scene as represented by the sensor data to determine one or more of a spatial or geometric differential between the previous scene as represented by the one or more alternative submaps and the contemporaneous scene as represented by the sensor data.
22. The one or more non-transitory computer-readable media of claim 21 , wherein:
the data representing the previous scene comprises data representing a point cloud representing the previous scene;
the sensor data representing the contemporaneous scene comprises data representing a point cloud representing the contemporaneous scene; and
the comparing comprises comparing the data representing the point cloud representing the previous scene with the data representing the point cloud representing the contemporaneous scene.
23. The one or more non-transitory computer-readable media of claim 21 , wherein:
the sensor data representing the contemporaneous scene comprises data generated based at least in part on data captured by one or more image sensors of the autonomous vehicle; and
the data representing the previous scene comprises data generated based at least in part on data captured by one or more corresponding image sensors.
24. The one or more non-transitory computer-readable media of claim 21 , wherein:
the sensor data representing the contemporaneous scene comprises data generated based at least in part on data captured by one or more light detection and ranging (LIDAR) sensors of the autonomous vehicle; and
the data representing the previous scene comprises data generated based at least in part on data captured by one or more corresponding LIDAR sensors.