Detection of misalignment hotspots for high definition maps for navigating autonomous vehicles

A system comprising:

• one or more processors to cause performance of operations comprising:
• controlling one or more navigation operations by a machine based at least on map data that is generated at least by:
• performing cross-track pairwise alignment between a first point cloud set including a first point cloud and a second point cloud set including a second point cloud, the cross-track pairwise alignment including determining respective cross-track alignments corresponding to one or more pairs of points respectively corresponding to the first point cloud and the second point cloud;
• performing a first single-track pairwise alignment between two or more first point clouds of the first point cloud set, the first single-track pairwise alignment being based at least on respective first single-track pairwise registrations between the two or more first point clouds of the first point cloud set;
• determining a first set of global poses for the first point cloud set based at least on a first combination of the cross-track pairwise alignment and the first single-track pairwise alignment;
• performing a second single-track pairwise alignment between two or more second point clouds of the second point cloud set, the second single-track pairwise alignment being based at least on respective second single-track pairwise registrations between the two or more second point clouds of the second point cloud set;
• determining a second set of global poses for the second point cloud set based at least on a second combination of the cross-track pairwise alignment and the second single-track pairwise alignment;
• and
• generating the map data to include the first point cloud set and the second point cloud set, the first point cloud set and the second point cloud set being aligned in the map data based at least on the first set of global poses and the second set of global poses.

A processor comprising processing circuitry to control one or more navigation operations by a machine based at least on map data, generation of the map data comprising:

• performing cross-track pairwise alignment based at least on first track data associated with a first machine navigating along a first path and second track data associated with a second machine navigating along a second path, the cross-track alignment being between a first point cloud set of the first track data and a second point cloud set of the second track data, the cross-track pairwise alignment including determining respective cross-track alignments corresponding to one or more pairs of points, one or more individual pairs of the one or more pairs of points including a respective first point of the first point cloud set and a respective second point of the second point cloud set;
• performing a single-track pairwise alignment between two or more first point clouds of the first point cloud set, the single-track pairwise alignment being based at least on respective first single-track pairwise registrations between the two or more first point clouds of the first point cloud set;
• determining a plurality of relative pairwise transformations respectively corresponding to the one or more pairs of points based at least on the cross-track pairwise alignment;
• aligning first sensor data of the first track data and second sensor data of the second track data based at least on a set of global poses determined based at least on a combination of the cross-track pairwise alignment and the single-track pairwise alignment the;
• and
• generating the map data using the first track data and the second track data as aligned.