Autonomous Driving Road Network Data Generation via Unit Lane Grouping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous driving systems lack an efficient method for generating accurate road network data, which is essential for route planning and navigation, as they struggle to effectively group and connect unit lanes based on varying lane boundaries and traveling directions.
Innovation Solution
A method and apparatus that utilize image analysis and graphing techniques to generate road network data by grouping unit lanes based on lane boundaries, traveling directions, and allowed travel areas, creating road graphs and connection information indicators to represent road networks accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If unit lanes are grouped based on lane boundaries to generate road network data, then the accuracy of road structure representation is improved, but the complexity of processing and analyzing image data increases
Solution Approach 1:
The patent segments the road network into multiple unit lanes, each representing a distinct drivable path. By dividing the complex road structure into manageable unit lanes with clear boundaries, the system can accurately represent road structures while processing each segment independently, reducing overall computational complexity.
Solution Approach 2:
The patent introduces connection information indicators as intermediaries to represent relationships between unit lanes. These indicators (such as connection types and travel directions) mediate the complex spatial relationships between lanes, allowing the system to accurately represent road networks without directly processing all possible lane interactions.
2Loss of information
If connection information indicators are generated to represent relationships between unit lanes, then the completeness of road network data is improved, but the amount of data to be processed and stored increases
Solution Approach 1:
The patent extracts only the essential connection information between unit lanes, such as connection types (merge, diverge, intersect) and travel directions. By taking out only the critical relationship data rather than processing all possible attributes, the system maintains completeness of road network representation while minimizing data volume.
Solution Approach 2:
The patent changes the representation parameters of connection information from detailed geometric descriptions to simplified categorical indicators (connection types, travel directions). This parameter transformation reduces data complexity while preserving the essential topological relationships needed for autonomous navigation.
3Reliability
If image analysis techniques are used to identify lane boundaries and group unit lanes, then the accuracy of autonomous driving navigation is improved, but the computational resources and time required increase
Solution Approach 1:
The patent performs preliminary image analysis to identify lane boundaries and group unit lanes before actual navigation tasks. By pre-processing and structuring the road network data into organized unit lanes with connection information, the system reduces computational requirements during real-time autonomous driving operations.
Solution Approach 2:
The patent replaces complex real-time image processing with pre-analyzed structured data. Instead of continuously analyzing raw images during navigation, the system substitutes mechanical image processing with queries to the pre-generated road network data structure, significantly reducing computational time while maintaining accuracy.
Data Source
AI summary
Provided are a method, an apparatus, and a computer program for generating road network data for an autonomous driving vehicle. The method of generating road network data for an autonomous driving vehicle, which is performed by a computing device, the method includes generating one or more roads by grouping a plurality of unit lanes, generating connection information about the one or more roads, and generating road network data including road graphs generated by graphing the one or more roads and reflecting the connection information on the graphed one or more roads.