Granted Patent B2

US 12,435,993 · App. 17/652,951 · Granted Oct 7, 2025

Inventor: Tero Juhani Keski-Valkama (Zurich, CH)

Assignee: HERE GLOBAL B.V.

G01C21/3848B60W60/001G01C21/3819G01C21/3822B60W2552/53B60W2554/20G01C21/3859

Abstract

Claims (57)

1. An apparatus comprising at least one processor and at least one non-transitory memory including computer program code instructions, the computer program code instructions configured to, when executed, cause the apparatus to at least:

receive a plurality of sequences of geospatial observations from discrete trajectories within a geographic area, wherein the geospatial observations comprise sensor data including observations of objects with one or more geo-positions and attributes received from one or more of image sensors, LiDAR (Light Distancing and Ranging) sensors, depth sensors, or sonic sensors;

refine a summary entity set representation comprising vector embeddings through iteration over the discrete trajectories;

determine a drive offset for each of the discrete trajectories based on a comparison of the summary entity set representation to a drive entity set for each of the discrete trajectories;

align the discrete trajectories to generate aligned geospatial observations based, at least in part, on the drive offset for a respective discrete trajectory;

concatenate the aligned geospatial observations;

process the concatenated, aligned geospatial observations using at least one Set Transformer;

generate, from the at least one Set Transformer, map geometries including objects from the concatenated, aligned geospatial observations;

update a map database to include the map geometries including the objects from the concatenated, aligned geospatial observations; and

provide for at least one of navigational assistance or at least semi-autonomous vehicle control based on the map geometries, wherein the map geometries including the objects from the concatenated, aligned geospatial observations in the map database are used with vehicle sensor data to align the vehicle within the geographic area.

2. The apparatus according to claim 1 , wherein causing the apparatus to determine the drive offset for each of the discrete trajectories based on the comparison of the summary entity set representation to the drive entity set for each of the discrete trajectories further comprises causing the apparatus to process an offset for each of the discrete trajectories through a pooling operation to obtain the drive offset for a respective drive.

3. The apparatus according to claim 1 , wherein causing the apparatus to align the discrete trajectories to generate the aligned geospatial observations based on the drive offset for the respective discrete trajectory comprises causing the apparatus to apply the drive offset as a geospatial offset to the respective discrete trajectory.

4. The apparatus according to claim 1 , wherein the plurality of sequences of geospatial observations from discrete trajectories include unique trajectory identifiers for each discrete trajectory, wherein causing the apparatus to concatenate the aligned geospatial observations comprises causing the apparatus to remove the unique trajectory identifiers associated with the geospatial observations.

5. The apparatus according to claim 1 , wherein causing the apparatus to determine a drive offset for each of the discrete trajectories based on the comparison of the summary entity set representation to the drive entity set for each of the discrete trajectories comprises causing the apparatus to determine at least one of an affine transformation or deformation for each of the discrete trajectories based on a comparison of the summary entity set representation to the drive entity set for each of the discrete trajectories.

6. The apparatus according to claim 1 , wherein causing the apparatus to refine the summary entity set representation through iteration over the discrete trajectories comprises causing the apparatus to:

initialize queries with a predefined number of values;

attend to the queries using a multi-head attention layer to the drive entity set for a single discrete trajectory; and

determine the summary entity set representation.

7. The apparatus according to claim 6 , wherein the summary entity set comprises cardinalities the same as the queries.

8. The apparatus according to claim 1 , wherein causing the apparatus to determine the drive offset for each of the discrete trajectories based on the comparison of the summary entity set representation to the drive entity set for each of the discrete trajectories comprises causing the apparatus to:

use the summary entity set representation in a multi-head attention layer to attend to the drive entity set for a single discrete trajectory; and

generate, from the multi-head attention layer, the drive offset.

9. The apparatus according to claim 1 , wherein the objects from the geospatial observations included in the map geometries comprise point objects and linear objects, wherein point objects comprise at least one of signs or poles, and wherein linear objects comprise at least one of road markings or road boundaries.

10. A method comprising:

receiving a plurality of sequences of geospatial observations from discrete trajectories;

refining a summary entity set representation comprising vector embeddings through iteration over the discrete trajectories within a geographic area, wherein the geospatial observations comprise sensor data including observations of objects with one or more geo-positions and attributes received from one or more of image sensors, LiDAR (Light Distancing and Ranging) sensors, depth sensors, or sonic sensors;

determining a drive offset for each of the discrete trajectories based on a comparison of the summary entity set representation to a drive entity set for each of the discrete trajectories;

aligning the discrete trajectories to generate aligned geospatial observations based, at least in part, on the drive offset for a respective discrete trajectory;

concatenating the aligned geospatial observations;

processing the concatenated, aligned geospatial observations using at least one Set Transformer;

generating, from the at least one Set Transformer, map geometries including objects from the concatenated, aligned geospatial observations;

updating a map database to include the map geometries including the objects from the concatenated, aligned geospatial observations; and

providing for at least one of navigational assistance or at least semi-autonomous vehicle control based on the map geometries, wherein the map geometries including the objects from the concatenated, aligned geospatial observations in the map database are used with vehicle sensor data to align the vehicle within the geographic area.

11. The method according to claim 10 , wherein determining the drive offset for each of the discrete trajectories based on the comparison of the summary entity set representation to the drive entity set for each of the discrete trajectories further comprises processing an offset for each of the discrete trajectories through a pooling operation to obtain the drive offset for a respective drive.

12. The method according to claim 10 , wherein aligning the discrete trajectories to generate the aligned geospatial observations based on the drive offset for the respective discrete trajectory comprises applying the drive offset as a geospatial offset to the respective discrete trajectory.

13. The method according to claim 10 , wherein the plurality of sequences of geospatial observations from discrete trajectories include unique trajectory identifiers for each discrete trajectory, wherein concatenating the aligned geospatial observations comprises removing the unique trajectory identifiers associated with the geospatial observations.

14. The method according to claim 10 , wherein determining a drive offset for each of the discrete trajectories based on the comparison of the summary entity set representation to the drive entity set for each of the discrete trajectories comprises determining at least one of an affine transformation or deformation for each of the discrete trajectories based on a comparison of the summary entity set representation to the drive entity set for each of the discrete trajectories.

15. The method according to claim 10 , wherein refining the summary entity set representation through iteration over the discrete trajectories comprises:

initializing queries with a predefined number of values;

attending to the queries using a multi-head attention layer to the drive entity set for a single discrete trajectory; and

determining the summary entity set representation.

16. The method according to claim 15 , wherein the summary entity set comprises cardinalities the same as the queries.

17. The method according to claim 10 , wherein determining the drive offset for each of the discrete trajectories based on the comparison of the summary entity set representation to the drive entity set for each of the discrete trajectories comprises:

using the summary entity set representation in a multi-head attention layer to attend to the drive entity set for a single discrete trajectory; and

generating, from the multi-head attention layer, the drive offset.

18. The method according to claim 10 , wherein the objects from the geospatial observations included in the map geometries comprise point objects and linear objects, wherein point objects comprise at least one of signs or poles, and wherein linear objects comprise at least one of road markings or road boundaries.

19. A computer program product comprising at least one non-transitory computer-readable storage medium having computer-executable program code instructions stored therein, the computer-executable program code instructions comprising program code instructions to:

receive a plurality of sequences of geospatial observations from discrete trajectories within a geographic area, wherein the geospatial observations comprise sensor data including observations of objects with one or more geo-positions and attributes received from one or more of image sensors, LiDAR (Light Distancing and Ranging) sensors, depth sensors, or sonic sensors;

refine a summary entity set representation comprising vector embeddings through iteration over the discrete trajectories;

determine a drive offset for each of the discrete trajectories based on a comparison of the summary entity set representation to a drive entity set for each of the discrete trajectories;

align the discrete trajectories to generate aligned geospatial observations based, at least in part, on the drive offset for a respective discrete trajectory;

concatenate the aligned geospatial observations;

process the concatenated, aligned geospatial observations using at least one Set Transformer;

generate, from the at least one Set Transformer, map geometries including objects from the concatenated, aligned geospatial observations;

update a map database to include the map geometries including the objects from the concatenated, aligned geospatial observations; and

provide for at least one of navigational assistance or at least semi-autonomous vehicle control based on the map geometries, wherein the map geometries including the objects from the concatenated, aligned geospatial observations in the map database are used with vehicle sensor data to align the vehicle within the geographic area.

20. The computer program product according to claim 19 , wherein the program code instructions to determine the drive offset for each of the discrete trajectories based on the comparison of the summary entity set representation to the drive entity set for each of the discrete trajectories further comprise program code instructions to process an offset for each of the discrete trajectories through a pooling operation to obtain the drive offset for a respective drive.

Assignments (1)

Continuity (1)

References Cited (41)