Driving data guided spatial planning

Patent metadata

Field	Value
Publication number	US-11946749-B2
Application number	US-202117200050-A
Country	US
Kind code	B2
Filing date	Mar 12, 2021
Priority date	Mar 12, 2021
Publication date	Apr 2, 2024
Grant date	Apr 2, 2024

Abstract

Among other things, techniques are described for driver data guided spatial planning. A spatial structure is generated comprising a plurality of nodes connected by edges. At least some of the nodes and edges represent a path to navigate a vehicle from a first point to a second point. Edges of the spatial structure are labeled as useful based on a distance metric. The spatial structure is pruned by removing one or more edges from the spatial structure according to a respective label of the edges, wherein an extent of the removal is based on a predetermined graph size, a predetermined performance, or any combinations thereof to obtain a pruned graph. A path from the first point to the second point on the pruned graph is identified and the vehicle is navigated in accordance with the path from the first point to the second point on the pruned graph.

First claim

What is claimed is:

1. A method comprising:
obtaining, using at least one processor, a spatial structure comprising a plurality of nodes connected by edges, wherein at least some of the nodes and edges represent a path to navigate a vehicle from a first point to a second point;
labeling, using the at least one processor, edges of the spatial structure as useful, based on, at least in part, a measure of how closely an edge replicates driving data samples from drive logs, wherein an edge is useful when a number of times the edge is found in the driving data samples satisfies a threshold and a sequence of edges is useful when a distance between the sequence of edges and the driving data samples satisfies a predetermined threshold;
pruning, using the at least one processor, the spatial structure by removing at least one edge from the spatial structure according to respective labels of the edges, wherein an extent of the removal is based on a predetermined graph size, a predetermined performance, or any combinations thereof to obtain a pruned graph;
adjusting the extent of the removal of the at least one edge from the spatial structure according to a measure of a coverage of paths in the drive logs by paths recreated on the pruned graph; and
navigating, using the at least one processor, the vehicle in accordance with a path from a first point to a second point representing a useful sequence of edges on the pruned graph.

2. The method of claim 1 , wherein obtaining the spatial structure comprises:
biasing an edge type of the spatial structure, wherein biasing the edge type of a graph layout generates a plurality of candidate graph layouts, and a graph layout of the plurality of candidate graph layouts is selected as the spatial structure based on an actual performance of the pruned graph.

3. The method of claim 1 , comprising:
labeling edges of the spatial structure as not useful, based on, at least in part, the measure of how closely an edge replicates driving data samples from drive logs; and
pruning the spatial structure by removing edges of the spatial structure labeled as not useful to obtain the pruned graph.

4. The method of claim 1 , wherein pruning comprises removing edges from the spatial structure according to at least one factor, wherein the at least one factor overrides a previous label of the edges.

5. The method of claim 1 , wherein pruning comprises removing the at least one edge from the spatial structure according to edge statistics.

6. The method of claim 1 , wherein an edge represents a series of adjacent locations without width forming a line between the first point and the second point.

7. The method of claim 1 , wherein an edge represents an area of locations forming a strip between the first point and the second point.

8. A non-transitory computer-readable storage medium comprising at least one program for execution by at least one processor of a first device, the at least one program including instructions which, when executed by the at least one processor, carry out a method comprising:
obtaining a spatial structure comprising a plurality of nodes connected by edges, wherein at least some of the nodes and edges represent a path to navigate a vehicle from a first point to a second point;
labeling edges of the spatial structure as useful, based on, at least in part, a measure of how closely an edge replicates driving data samples from drive logs, wherein an edge is useful when a number of times the edge is found in the driving data samples satisfies a threshold and a sequence of edges is useful when a distance between the sequence of edges and the driving data samples satisfies a predetermined threshold;
pruning the spatial structure by removing at least one edge from the spatial structure according to respective labels of the edges, wherein an extent of the removal is based on a predetermined graph size, a predetermined performance, or any combinations thereof to obtain a pruned graph;
adjusting the extent of the removal of the at least one edge from the spatial structure according to a measure of a coverage of paths in the drive logs by paths recreated on the pruned graph; and
navigating the vehicle in accordance with a path from a first point to a second point representing a useful sequence of edges on the pruned graph.

9. The computer-readable storage medium of claim 8 , wherein obtaining the spatial structure comprises biasing an edge type of the spatial structure, wherein biasing the edge type of a graph layout generates a plurality of candidate graph layouts, and a graph layout of the plurality of candidate graph layouts is selected as the spatial structure based on an actual performance of the pruned graph.

10. A method, comprising:
evaluating, using at least one processor, a pruned graph comprising a plurality of nodes connected by edges, wherein the edges of the pruned graph are labeled as useful based on a distance and driving data samples from drive logs and an extent of removal of at least one edge from a spatial structure to generate the pruned graph is based on a measure of a coverage of paths in the drive logs by paths recreated on the pruned graph;
iteratively adjusting, using the at least one processor, edges selected for removal from the pruned graph based on a respective performance of the pruned graph;
selecting, using the at least one processor, a remaining pruned graph with a highest performance, wherein the respective performance of the pruned graph is evaluated at the iterative adjustments; and
navigating, using the at least one processor, a vehicle in accordance with a path from a first point to a second point on the selected, remaining pruned graph.

11. The method of claim 10 , comprising:
generating a spatial structure comprising a plurality of nodes connected by edges, wherein a preferred edge type is biased in accordance with statistics and an actual performance of the pruned graph;
labeling the edges as useful or not useful based on a distance and drive logs;
pruning the spatial structure by removing edges labeled as not useful to obtain a final pruned graph; and
navigating the vehicle in accordance with the path from the first point to the second point on the final pruned graph.

12. The method of claim 10 , wherein labeling edges of a spatial structure as useful comprises calculating a respective average distance for the edges of the spatial structure and corresponding portions of trajectories in the drive logs, wherein edges with a respective average distance that satisfies a first predetermined threshold are labeled as useful.

13. The method of claim 10 , wherein iteratively adjusting edges removed from the pruned graph based on the respective performance of the pruned graph comprises:
removing at least one edge from the pruned graph; and
reevaluating the respective performance of the remaining pruned graph.

14. The method of claim 10 , comprising removing edges from the pruned graph to satisfy a predetermined graph size, wherein the predetermined graph size is measured by a number of edges in a graph, a storage requirement of the graph, or any combinations thereof.

15. The method of claim 10 , wherein the path from the first point to the second point on the pruned graph is a lowest cost path, wherein a cost of a path is a value that represents resources expended when navigating the vehicle on a respective path.

16. The method of claim 10 , wherein performance of the pruned graph is an accuracy of edges in replicating drive log trajectories.

Assignees

• Motional Ad Llc

Inventors

• Cserna Bence
• Gu Tianyi
• Wolff Eric
• Grigore Elena Corina
• Shrestha Mochan

Classifications

• B60W2556/50

  of positioning data, e.g. GPS [Global Positioning System] data · CPC title

• G01C21/3407 Primary

  specially adapted for specific applications · CPC title

• G01C21/32 Primary

  Structuring or formatting of map data · CPC title

• B60W40/06

  Road conditions · CPC title

• B60W60/001 Primary

  Planning or execution of driving tasks · CPC title