System and method for generating shard data based on semantic map sharding

Patent metadata

Field	Value
Publication number	US-11526535-B2
Application number	US-202117219773-A
Country	US
Kind code	B2
Filing date	Mar 31, 2021
Priority date	Mar 31, 2021
Publication date	Dec 13, 2022
Grant date	Dec 13, 2022

Abstract

A system and method for generating shard data based on semantic sharding is proposed. Shard data is generated by a map sharder that is configured to generate the shard data based on semantic objects in a geographical area and a definition of sections. The sections cover parts of the geographical area. The map sharder searches, for each of the sections, for semantic objects that are located at least partly in a section and stores found objects in a shard data entry of the shard data.

First claim

What is claimed is:

1. A system for generating shard data, the system comprising:
one or more memories having instructions stored thereon;
one or more database storage devices;
one or more processors to execute the instructions;
a map sharder, encoded in the instructions, to:
receive a first data input comprising semantic objects in a geographical area and a second data input comprising a definition of sections, wherein the sections cover parts of the geographical area;
find, for each of the sections, semantic objects that are located at least partly in a section;
apply consistency rules to the found semantic objects for each of the sections, wherein:
the consistency rules define which semantic objects are to be combined into a single shard and which semantic objects are to be stored in separate shards, and
the consistency rules comprise: closely related primary semantic objects including lanes, boundaries, and intersections, are combined into a single shard based on lane to boundary relationships and intersection identifications, and a given semantic object is contained in one and only shard;

after consistency rules are applied, link found objects for each of the sections that obey the consistency rules to respective shards in a form of shard data entries;
storing the shard data entries in the one or more database storage devices; and
compute data for bounding boxes based on the linked found objects of each shard; and

an uploader, encoded in the instructions, to upload files to clients created based on the shard data entries and the computed data for the bounding boxes, wherein clients include one or more of an autonomous vehicle, a vehicle, and a computer environment that tests software for vehicles.

2. The system according to claim 1 , wherein the consistency rules further include:
objects related to one or more primary objects which are to be stored in the same shard data entry as the related primary object.

3. The system according to claim 1 , wherein the consistency rules further include:
objects with large geometries or no relations to other objects that cannot be reasonably placed in one of the sections, which are to be stored in separate shard data entries.

4. The system according to claim 1 , wherein the shard data comprises:
a first database in the one or more database storage devices to store metadata for each shard, the metadata for a given shard comprising an identifier of the given shard, computed data representative of the bounding box for the given shard, and a subtype of the semantic objects belonging to the shard; and
a second database in the one or more storage devices to store the shard data entries, where each of the shard data entries relates semantic objects belonging to the given shard.

5. The system according to claim 1 , wherein the consistency rules further comprise:
objects with geometries, which are to be added to a shard entry of a section that contains most of that geometry.

6. The system according to claim 1 , wherein the consistency rules further include:
objects closely related to an object with geometry, which are to be stored into a shard data entry together with the geometric object.

7. The system according to claim 1 , further comprising:
a database exporter configured to generate the files from the shard data entries, wherein one file comprises data for one shard data entry; and
a file storage system configured to store the files.

8. The system according to claim 7 , wherein the database exporter is further configured to generate a metadata file comprising the computed data of the bounding boxes for the shard data entries, and wherein the file storage system is further configured to store the metadata file.

9. The system according to claim 8 , wherein the uploader is to upload the files and the metadata file to a cloud storage.

10. The system according to claim 7 , wherein the database exporter is further configured to generate monolithic map data, and wherein the file storage system is further configured to store the monolithic map data.

11. The system according to claim 1 , wherein the consistency rules further include:
relationships between objects in different shards data entries, which are to be stored with one of the objects.

12. A method for generating shard data, comprising:
receiving, in a map sharding function encoded in instructions stored in one or more memories, wherein the instructions are executable by one or more processors, and a first data input comprises semantic objects in a geographical area;
receiving, in the map sharding function, a second data input comprising a definition of sections, wherein the sections cover parts of the geographical area;
finding, by the map sharding function, for each of the sections, semantic objects that are located at least partly in a section;
applying consistency rules to the found semantic objects for each of the sections, wherein:
the consistency rules define which semantic objects are to be combined into a single shard and which semantic objects are to be stored in separate shards, and
the consistency rules comprise: closely related primary semantic objects including lanes, boundaries, and intersections, are combined into a single shard based on lane to boundary relationships and intersection identifications and a given semantic object is contained in one and only shard;

after consistency rules are a lied linking found objects for each of the sections that obey the consistency rules to respective shards in a form of shard data entries;
storing the shard data entries in one or more database storage devices;
computing data for bounding boxes based on the linked found objects of each shard; and
uploading, by an uploader function encoded in the instructions, files to clients created based on the shard data entries and the computed data for the bounding boxes, wherein clients include one or more of an autonomous vehicle, a vehicle, and a computer environment that tests software for vehicles.

13. The method according to claim 12 ,
wherein the consistency rules further include:
objects related to one or more primary objects which are to be stored in the same shard data entry as the related primary object.

14. The method according to claim 12 , wherein the computed data for the bounding boxes is stored as metadata separate from the shard data entries.

15. The method according to claim 12 ,
wherein the consistency rules further include:
objects with large geometries or no relations to other objects that cannot be reasonably placed in one of the sections, which are to be stored in separate shard data entries.

16. The method according to claim 12 ,
wherein the consistency rules further include:
objects with geometries which are to be added to a shard entry of a section that contains most of that geometry.

17. The method according to claim 12 , wherein the consistency rules further include:
objects closely related to an object with geometry, which are to be stored into a shard data entry together with the geometric object; and/or
relationships between objects in different shards data entries, which are to be stored with one of the objects.

18. One or more non-transitory computer-readable storage media comprising instruction for execution which, when executed by a processor, are operable to perform operations for generating files, the operations comprising:
inputting a first data input comprising semantic objects in a geographical area;
inputting a second data input comprising a definition of sections, where

Assignees

• Gm Cruise Holdings Llc

Inventors

• Chernov Mikhail
• Sun Xiaojun

Classifications

• G01C21/3896 Primary

  Transmission of map data from central databases · CPC title

• G01C21/387

  Organisation of map data, e.g. version management or database structures · CPC title

• G06F16/13

  File access structures, e.g. distributed indices (arrangements of input from, or output to, record carriers G06F3/06) · CPC title

• G06F16/29

  Geographical information databases · CPC title

• G06F16/285 Primary

  Clustering or classification · CPC title