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

Patent metadata

Field	Value
Publication number	US-11280609-B2
Application number	US-201715857609-A
Country	US
Kind code	B2
Filing date	Dec 28, 2017
Priority date	Dec 30, 2016
Publication date	Mar 22, 2022
Grant date	Mar 22, 2022

Abstract

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.

First claim

What is claimed is:

1. A method for determining hotspots in alignment data for generating high definition maps for use in driving of autonomous vehicles, the method comprising:
receiving sensor data captured by a plurality of vehicles driving through a path in a geographical region, wherein the sensor data includes scan data captured by a respective LIDAR mounted on each of two or more vehicles of the plurality of vehicles;
combining the received sensor data to generate a three dimensional representation of the geographical region;
determining a surface and a measure of thickness corresponding to the surface from the three dimensional representation of the geographical region;
determining a measure of likelihood of misalignment of at least a portion of the identified surface as a value proportionate to the measure of thickness;
marking, as a misalignment hotspot location, an area identified within the three dimensional representation that includes the portion of the identified surface, the area being identified as the misalignment hotspot location based on the determined measure of likelihood of misalignment;
obtaining a modification to the three dimensional representation in which the three dimensional representation has been modified in response to the marking of the area;
generating high definition map data based on the modified three dimensional representation;
providing the high definition map data to a vehicle; and
causing the performance of one or more driving operations by the vehicle using the high definition map data.

2. The method of claim 1 , wherein marking the area includes providing a heat map identifying the area.

3. The method of claim 1 , wherein the three dimensional representation of the geographical region is a point cloud.

4. The method of claim 1 , wherein the sensor data further comprises images captured by a respective camera mounted on each of the one or more vehicles.

5. The method of claim 1 , wherein determining the measure of thickness corresponding to the surface comprises:
determining a normal vector pointing in a direction normal to the identified surface;
selecting a point on the identified surface;
identifying a cluster of points related to the selected point in the three dimensional representation along the normal direction and a direction opposite the normal direction; and
determining a maximum distance between the points in the cluster as the measure of thickness corresponding to the identified surface.

6. The method of claim 5 , wherein the identified surface represents ground and the normal vector points in a vertical direction.

7. The method of claim 5 , wherein the identified surface represents a wall and the normal vector is in a horizontal plane.

8. The method of claim 7 , wherein identifying the cluster of points comprises excluding a point if the point has a vertical coordinate that is different from the selected point on the identified surface.

9. The method of claim 7 , wherein identifying the cluster of points comprises excluding a point if a vector direction from the point to the selected point on the identified surface is not parallel to the normal vector.

10. The method of claim 1 , further comprising:
identifying a first point and a second point along the path in the geographical region;
determining a first measure of distance between the first point and the second point representing a distance traversed along the path to get from the first point to the second point;
determining a second measure of distance based on a geodesic distance between the first point and the second point; and
determining, as the measure of likelihood of misalignment, a measure of misalignment probability associated with the first point and the second point as a value proportionate to the difference between the first measure of distance and the second measure of distance.

11. One or more non-transitory computer readable storage media storing instructions that, in response to being executed by one or more processors, cause a system to perform operations, the operations comprising:
receiving sensor data captured by a plurality of vehicles driving through a path in a geographical region, wherein the sensor data includes scan data captured by a respective LIDAR mounted on each of one or more vehicles of the plurality of vehicles;
combining the received sensor data to generate a three dimensional representation of the geographical region, wherein high definition map data is generated based on the three dimensional representation;
determining a surface and a measure of thickness corresponding to the surface from the three dimensional representation of the geographical region;
determining a measure of likelihood of misalignment of at least a portion of the identified surface as a value proportionate to the measure of thickness;
marking, as a misalignment hotspot location, an area identified within the three dimensional representation that includes the portion of the identified surface, the area being identified as the misalignment hotspot location based on the determined measure of likelihood of misalignment; and
obtaining a modification to the high definition map data in which the high definition map data has been modified in response to the marking of the area.

12. The non-transitory computer readable storage media of claim 11 , wherein determining the measure of thickness corresponding to the surface comprises:
determining a normal vector pointing in a direction normal to the identified surface;
selecting a point on the identified surface;
identifying a cluster of points related to the selected point in the three dimensional representation along the normal direction and a direction opposite the normal direction; and
determining a maximum distance between the points in the cluster as the measure of thickness corresponding to the identified surface.

13. The non-transitory computer readable storage media of claim 11 , wherein the identified surface represents ground and the normal vector points in a vertical direction.

14. The non-transitory computer readable storage media of claim 11 , wherein the identified surface represents a wall and the normal vector is in a horizontal plane.

15. The non-transitory computer readable storage media of claim 11 , wherein the operations further comprise:
identifying a first point and a second point along the path in the geographical region;
determining a first measure of distance between the first point and the second point representing a distance traversed along the path to get from the first point to the second point;
determining a second measure of distance based on a geodesic distance between the first point and the second point; and
determining, as the measure of likelihood of misalignment, a measure of misalignment probability associated with the first point and the second point as a value proportionate to the difference between the first measure of distance and the second measure of distance.

16. A computer system comprising:
an electronic processor; and
one or more non-transitory computer readable storage media storing instructions that, in response to being executed by the electronic processor, cause the system to perform operations, the operations comprising:
receiving sensor data captured by a plurality of vehicles driving through a path in a geographical region, wherein the sensor data includes scan data captured by a respective LIDAR mounted on each of one or more vehicles of the plurality of vehicles;
combining the received sensor data to generate a three dimensional representation of t

Assignees

• Deepmap Inc
• Nvidia Corp

Inventors

• Chen Chen
• Wheeler Mark Damon
• Zou Liang

Classifications

• G01C11/12 Primary

  the pictures being supported in the same relative position as when they were taken · CPC title

• G06T17/05 Primary

  Geographic models · CPC title

• G06V10/462

  Salient features, e.g. scale invariant feature transforms [SIFT] · CPC title

• G06V10/758

  Involving statistics of pixels or of feature values, e.g. histogram matching · CPC title

• G01C21/3867

  Geometry of map features, e.g. shape points, polygons or for simplified maps · CPC title