Graph Neural Road Parsing for Compact Drivable Path Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in navigating due to the vast amount of data required for processing and storing traditional maps, which can limit their navigation capabilities and pose storage and update difficulties.
Innovation Solution
A system and method using a graph neural network to predict drivable paths by generating topographical representations of road segments and encoding actual vehicle trajectories, allowing for efficient storage and update of map data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional mapping technology is used for autonomous vehicle navigation, then navigation capabilities are provided, but data storage requirements and update complexity increase significantly
Solution Approach 1:
The patent extracts only the essential drivable path information from complete traditional maps. Instead of storing and processing entire map datasets, the system identifies and stores only the critical path segments that vehicles actually need to navigate, significantly reducing data volume while maintaining navigation functionality
Solution Approach 2:
The patent segments the continuous map data into discrete drivable path segments. By dividing the map into individual path segments that can be independently stored and processed, the system reduces overall data requirements and enables more efficient storage and update mechanisms
2Reliability
If traditional mapping technology is used for autonomous vehicle navigation, then navigation capabilities are provided, but map update difficulty and frequency increase
Solution Approach 1:
The system extracts only the necessary path segment information from complete maps, which simplifies the update process. When map updates are needed, only the affected path segments need to be updated rather than entire map datasets, reducing update complexity and frequency
Solution Approach 2:
The patent implements a dynamic map representation where only changed path segments are updated. This allows the navigation system to adapt to map changes efficiently by updating only the necessary portions rather than performing full map updates, reducing both complexity and resource requirements
3Measurement precision
If vast volumes of navigation data are processed and stored, then navigation accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The patent extracts only the essential path segment data needed for accurate location identification, eliminating unnecessary map details. This reduction in data volume directly decreases processing time while maintaining the precision needed for accurate vehicle location determination
Solution Approach 2:
By segmenting map data into discrete path segments, the system enables more efficient processing. The autonomous vehicle can quickly search and match its current position against segmented path segments rather than processing continuous large-scale map data, improving both speed and accuracy
Data Source
AI summary
Systems and methods for predicting drivable paths relative to road segments are disclosed. In one implementation, a system includes a processor programmed to access topographical information associated with a road segment; generate a topographical representation of the road segment based on the topographical information; input the topographical representation of the road segment to a trained model, wherein the trained model includes a graph neural network and is configured to predict at least one drivable path relative to the road segment based on the topographical representation of the road segment; receive, from the trained model, information identifying the drivable path; and store the information identifying the drivable path in a map.