Distributed processing of pose graphs for generating high definition maps for navigating autonomous vehicles
A high-definition map system receives sensor data from vehicles travelling along routes and combines the data to generate a high definition map for use in driving vehicles, for example, for guiding autonomous vehicles. A pose graph is built from the collected data, each pose representing location and orientation of a vehicle. The pose graph is optimized to minimize constraints between poses. Points associated with surface are assigned a confidence measure determined using a measure of hardness/softness of the surface. A machine-learning-based result filter detects bad alignment results and prevents them from being entered in the subsequent global pose optimization. The alignment framework is parallelizable for execution using a parallel/distributed architecture. Alignment hot spots are detected for further verification and improvement. The system supports incremental updates, thereby allowing refinements of subgraphs for incrementally improving the high-definition map for keeping it up to date.
1. A method for distributed processing of pose graphs for generating high-definition maps, the method comprising:
receiving sensor data captured by a plurality of vehicles driving through a path in a geographical region;
generating a pose graph, wherein each node of the pose graph represents a pose of a vehicle, the pose comprising a location and orientation of the vehicle, and wherein each edge between a pair of nodes represents a transformation between nodes of the pair of nodes;
dividing the pose graph into a plurality of subgraphs, each subgraph including a set of core nodes and a set of buffer nodes;
iteratively performing the steps comprising:
for each subgraph, keeping values of boundary nodes fixed and optimizing the subgraph, and
updating all node poses using core node poses of all subgraphs;
generating a high-definition map based on a point cloud representation; and
sending the high-definition map to one or more autonomous vehicles for navigating the autonomous vehicles in the geographical region.
2. The method of claim 1 , wherein the sensor data comprises data collected by LIDAR, data collected by global positioning system (GPS), and data collected by inertial measurement unit (IMU).
3. The method of claim 1 , wherein repeatedly performing the steps further comprises, determining whether there are changes in boundary nodes as a result of updating all node poses, and stopping the iterations responsive to determining that the changes to boundary nodes are below a threshold value.
4. The method of claim 1 , wherein the subgraphs are distributed among a plurality of processors for distributed execution.
5. The method of claim 1 , wherein dividing the pose graph into a plurality of subgraphs comprises:
determining bounding boxes based on latitude and longitude values; and
assigning all poses within a bounding box to a subgraph.
6. The method of claim 1 , wherein dividing the pose graph into a plurality of subgraphs comprises dividing the pose graph along a road to obtain subgraphs.
7. The method of claim 1 , wherein dividing the pose graph into a plurality of subgraphs comprises dividing the pose graph along a junction in roads to obtain subgraphs.
8. The method of claim 1 , wherein dividing the pose graph into a plurality of subgraphs comprises dividing the pose graph with the objective of minimizing the number of boundary nodes.
9. The method of claim 1 , wherein dividing the pose graph into a plurality of subgraphs comprises identifying portions of a geographical region that have large number of samples returned by vehicles and dividing the pose graph into subgraphs such that a boundary of a subgraph passes through an identified portion of the geographical region.
10. A non-transitory computer readable storage medium storing instructions for:
receiving sensor data captured by a plurality of vehicles driving through a path in a geographical region;
generating a pose graph, wherein each node of the pose graph represents a pose of a vehicle, the pose comprising a location and orientation of the vehicle, and wherein each edge between a pair of nodes represents a transformation between nodes of the pair of nodes;
dividing the pose graph into a plurality of subgraphs, each subgraph including a set of core nodes and a set of buffer nodes;
iteratively performing the steps comprising:
for each subgraph, keeping values of boundary nodes fixed and optimizing the subgraph, and
updating all node poses using core node poses of all subgraphs;
generating a high-definition map based on a point cloud representation; and
sending the high-definition map to one or more autonomous vehicles for navigating the autonomous vehicles in the geographical region.
11. The non-transitory computer readable storage medium of claim 10 , wherein the sensor data comprises data collected by LIDAR, data collected by global positioning system (GPS), and data collected by inertial measurement unit (IMU).
12. The non-transitory computer readable storage medium of claim 10 , wherein repeatedly performing the steps further comprises, determining whether there are changes in boundary nodes as a result of updating all node poses, and stopping the iterations responsive to determining that the changes to boundary nodes are below a threshold value.
13. The non-transitory computer readable storage medium of claim 10 , wherein the subgraphs are distributed among a plurality of processors for distributed execution.
14. The non-transitory computer readable storage medium of claim 10 , wherein instructions for dividing the pose graph into a plurality of subgraphs comprise instructions for:
determining bounding boxes based on latitude and longitude values; and
assigning all poses within a bounding box to a subgraph.
15. The non-transitory computer readable storage medium of claim 10 , wherein dividing the pose graph into a plurality of subgraphs comprises dividing the pose graph along a road to obtain subgraphs.
16. The non-transitory computer readable storage medium of claim 10 , wherein dividing the pose graph into a plurality of subgraphs comprises dividing the pose graph along a junction in roads to obtain subgraphs.
17. The non-transitory computer readable storage medium of claim 10 , wherein dividing the pose graph into a plurality of subgraphs comprises dividing the pose graph with the objective of minimizing the number of boundary nodes.
18. The non-transitory computer readable storage medium of claim 10 , wherein dividing the pose graph into a plurality of subgraphs comprises identifying portions of a geographical region that have large number of samples returned by vehicles and dividing the pose graph into subgraphs such that a boundary of a subgraph passes through an identified portion of the geographical region.
19. A computer system comprising:
an electronic processor; and
a non-transitory computer readable storage medium storing instructions executable by the electronic processor, the instructions for:
receiving sensor data captured by a plurality of vehicles driving through a path in a geographical region;
generating a pose graph, wherein each node of the pose graph represents a pose of a vehicle, the pose comprising a location and orientation of the vehicle, and wherein each edge between a pair of nodes represents a transformation between nodes of the pair of nodes;
dividing the pose graph into a plurality of subgraphs, each subgraph including a set of core nodes and a set of buffer nodes;
iteratively performing the steps comprising:
for each subgraph, keeping values of boundary nodes fixed and optimizing the subgraph, and
updating all node poses using core node poses of all subgraphs;
generating a high-definition map based on a point cloud representation; and
sending the high-definition map to one or more autonomous vehicles for navigating the autonomous vehicles in the geographical region.
20. The computer system of claim 19 , wherein instructions for dividing the pose graph into a plurality of subgraphs comprise instructions for:
determining bounding boxes based on latitude and longitude values; and
assigning all poses within a bounding box to a subgraph.