Overlapping Road Vertical Order Using Non-Self-Overlapping Polygons
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to accurately determine the vertical order of overlapping roads in 2D maps, leading to challenges in navigation and processing sensor data due to ambiguity and difficulty in discerning which road a vehicle is situated on, especially in stacked road scenarios where the order can change.
Innovation Solution
A system and method that utilize vehicle sensor data to generate non-self-overlapping polygons and determine vertical order by analyzing elevation differences and proximity, forming a drivability graph and separation graph to partition roads into layers without self-overlap, and assign a vertical order based on elevation indicators.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional 2D mapping methods are used to represent overlapping roads, then the map structure remains simple, but the ability to determine vertical order and distinguish between stacked roads deteriorates
Solution Approach 1:
The patent introduces a vertical dimension (elevation level) to the traditional 2D map representation. By assigning elevation levels to road segments and using 3D bounding boxes with height information, the system can distinguish between overlapping roads that appear identical in 2D space. This dimensional expansion resolves the ambiguity in determining which road a vehicle is on when roads stack vertically.
Solution Approach 2:
The patent segments the road network into discrete road segments with associated elevation levels and 3D bounding boxes. Each overlapping road is divided into separate segments that can be individually characterized with vertical position information. This segmentation allows the system to process and distinguish between multiple roads at the same 2D location by treating them as separate entities with different vertical attributes.
2Reliability
If sensor data is processed without vertical order information, then processing speed remains fast, but navigation accuracy deteriorates due to ambiguity in identifying the correct road
Solution Approach 1:
The patent introduces 3D bounding boxes and elevation levels as intermediary data structures that mediate between raw sensor data and navigation decisions. These intermediaries carry vertical position information that resolves ambiguity without requiring complex processing algorithms. The bounding boxes serve as a bridge that encodes spatial relationships in a format that can be efficiently queried and used for accurate road identification.
3Measurement precision
If detailed 3D information is stored for all road segments, then vertical order determination improves, but computational and storage requirements increase
Solution Approach 1:
The patent applies 3D bounding box information and elevation levels selectively to road segments that are part of overlapping or stacked road configurations. Not all road segments require full 3D characterization - only those involved in vertical overlaps need the additional data. This local application of detailed information reduces overall storage requirements while maintaining accuracy where it is most needed.
Data Source
AI summary
Systems, methods, and other embodiments described herein relate to determining a vertical order of overlapping road segments. In one embodiment, a method includes generating connections between centroids based on at least an elevation difference, proximity, and forming a non-self-overlapping polygon. The centroids are vehicle positions along one or more roads. The non-self-overlapping polygon has sides that match at least a portion of the one or more roads and do not overlap. The method includes determining a vertical order of the one or more roads based on at least the non-self-overlapping polygon and an indicator indicating an elevation of the non-self-overlapping polygon.