Method and apparatus for context based map data retrieval

Patent metadata

Field	Value
Publication number	US-11263245-B2
Application number	US-201816175394-A
Country	US
Kind code	B2
Filing date	Oct 30, 2018
Priority date	Oct 30, 2018
Publication date	Mar 1, 2022
Grant date	Mar 1, 2022

Abstract

An approach is provided for storing and retrieving map data using contextual information priors. The approach involves, for example, processing contextual information to determine a restricted range of location information relevant to at least one query. The approach also involves processing sensor data received from at least one sensor, the sensor data collected at at least one query location, to determine semantic information. The approach further involves filtering the map data based, at least in part, on the restricted range of location information relevant to the at least one query, the semantic information, or a combination thereof.

First claim

What is claimed is:

1. A computer-implemented method for use in training a plurality of neural networks for map data retrieval, the method comprising:
processing contextual information to determine a restricted range of location information relevant to at least one query;
processing sensor data received from at least one sensor, the sensor data collected at at least one query location, to determine semantic information;
filtering the map data based, at least in part, on the restricted range of location information relevant to the at least one query, the semantic information, or a combination thereof; and
retrieving only the filtered map data from a geographic database in response to the at least one query.

2. The method of claim 1 , further comprising:
processing vehicle position and/or heading data received from the at least one sensor to determine a geometric context for a driving direction of at least one vehicle; and
filtering the map data based, at least in part, on the geometric context for a driving direction of the at least one vehicle.

3. The method of claim 2 , further comprising:
determining the semantic information by performing semantic segmentation of the sensor data to identify and segment one or more relevant regions in the sensor data.

4. The method of claim 3 , further comprising:
determining the geometric context for a driving direction of the at least one vehicle based, at least in part, on the semantic segmentation.

5. The method of claim 1 , wherein the plurality of neural networks regresses the at least one query location on a mapping platform to separate unfiltered and filtered map data.

6. The method of claim 1 , wherein the semantic information comprises one or more road signs, one or more lane lines, terrain features, drivable surfaces, or a combination thereof relevant to the at least one query location.

7. The method of claim 1 , wherein the sensor data comprises visual data, aural data, light detection and ranging (LIDAR) data, or a combination thereof at least one query location.

8. The method of claim 1 , wherein the contextual information comprises sensor information, temporal information, vehicle position information, seasonal information, temperature information, or a combination thereof.

9. The method of claim 1 , wherein a neural network is used to process the contextual information to determine the restricted range of location information.

10. The method of claim 9 , wherein the neural network is trained using the training data set comprising the filtered map data.

11. An apparatus for in training a plurality of neural networks for map data retrieval, comprising:
at least one processor; and
at least one memory including computer program code for one or more programs,
the computer program code executed by the at least one processor, cause the apparatus to perform at least the following,
process contextual information to determine a restricted range of location information relevant to at least one query;
process sensor data received from at least one sensor, the sensor data collected at at least one query location, to determine semantic information;
filter the map data based, at least in part, on the restricted range of location information relevant to the at least one query, the semantic information, or a combination thereof; and
retrieve only the filtered map data from a geographic database in response to the at least one query.

12. The apparatus of claim 11 , wherein the apparatus is further caused to:
process vehicle position and/or heading data received from the at least one sensor to determine a geometric context for a driving direction of at least one vehicle; and
filter the map data based, at least in part, on the geometric context for a driving direction of the at least one vehicle.

13. The apparatus of claim 12 , further comprising:
determine the semantic information by performing semantic segmentation of the sensor data to identify and segment one or more relevant regions in the sensor data.

14. The apparatus of claim 13 , wherein the apparatus is further caused to:
determine the geometric context for a driving direction of the at least one vehicle based, at least in part, on the semantic segmentation.

15. The apparatus of claim 11 , wherein the plurality of neural networks regresses the at least one query location on a mapping platform to separate unfiltered and filtered map data.

16. The apparatus of claim 11 , wherein the semantic information comprises one or more road signs, one or more lane lines, terrain features, drivable surfaces, or a combination thereof relevant to the at least one query location.

17. The apparatus of claim 11 , wherein the sensor data comprises visual data, aural data, light detection and ranging (LIDAR) data, or a combination thereof at least one query location.

18. A non-transitory computer-readable storage medium for training a plurality of neural networks for map data retrieval, carrying one or more sequences of one or more instructions which, when executed by one or more processors, cause an apparatus to at least perform the following steps:
processing contextual information to determine a restricted range of location information relevant to at least one query;
processing sensor data received from at least one sensor, the sensor data collected at at least one query location, to determine semantic information;
filtering the map data based, at least in part, on the restricted range of location information relevant to the at least one query location, the semantic information, or a combination thereof; and
retrieving only the filtered map data from a geographic database in response to the at least one query.

19. The non-transitory computer-readable storage medium of claim 18 , further comprising:
processing vehicle position and/or heading data received from the at least one sensor to determine a geometric context for a driving direction of at least one vehicle; and
filtering the map data based, at least in part, on the geometric context for a driving direction of the at least one vehicle.

20. The non-transitory computer-readable storage medium of claim 19 , further comprising:
determining the semantic information by performing semantic segmentation of the sensor data to identify and segment one or more relevant regions in the sensor data.

Assignees

• Here Global Bv

Inventors

• Viswanathan Anirudh

Classifications

• G06F18/2135

  based on approximation criteria, e.g. principal component analysis · CPC title

• G06F18/24

  Classification techniques · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/0455

  Auto-encoder networks; Encoder-decoder networks · CPC title