Synthetic HD Map Geometry from SD Road Data for Autonomous Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Generating high-definition (HD) maps for autonomous or semi-autonomous systems is costly and time-consuming, and existing simulated maps lack the complexity and detailed information required for real-world applications.
Innovation Solution
Utilizing standard definition (SD) map data, systems and methods apply rule-based approaches to calculate geometric features, randomly assign characteristics, and apply logical rules to generate HD map data, including lane assignments and traffic management components, thereby reducing computational costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high definition (HD) maps are generated using traditional methods, then detailed and accurate road and traffic information is obtained, but the process is costly and time-consuming
Solution Approach 1:
The system performs preliminary processing by extracting geometric features from existing standard definition (SD) map data before generating HD map data. This preliminary extraction of road segments, intersections, and geometric properties enables subsequent rapid generation of detailed map information without starting from scratch, thus reducing overall generation time while maintaining accuracy
Solution Approach 2:
The system creates synthetic HD map data by copying and transforming existing SD map data structures. It replicates road geometries, intersections, and traffic components from SD sources, then enhances them with additional detailed information through rule-based generation, effectively creating high-definition versions of existing low-definition maps
2Manufacturing precision
If high definition (HD) maps are generated using traditional methods, then detailed and accurate road and traffic information is obtained, but computational costs increase
Solution Approach 1:
The map generation process is segmented into distinct rule-based modules: extracting geometric features from SD data, generating road segments, creating intersections, adding traffic components, and assigning lane configurations. Each segment operates independently with specific rules, reducing overall computational complexity compared to monolithic HD map generation approaches
Solution Approach 2:
The system uses existing SD map data as self-service input, automatically extracting necessary geometric features and road structures without requiring additional expensive data collection missions. The SD map data itself provides the foundation for generating HD maps, eliminating the need for separate high-cost data acquisition processes
3Ease of manufacture
If simulated maps are used for training autonomous systems, then training can be performed, but the maps lack the complexity of real-world applications
Solution Approach 1:
The system transforms SD map data into HD map data by changing key parameters including adding detailed lane assignments, traffic light configurations, crosswalk locations, and intersection geometries. This parameter enhancement maintains the underlying road structure while significantly increasing map complexity and detail to match real-world conditions needed for robust autonomous system training
Data Source
AI summary
Approaches presented herein provide for high definition (HD) map data generation from standard definition (SD) map data. Geometric properties may be extracted from the SD map data to form a geometric representation. The geometric representation may be processed to determine locations to insert traffic management components to simulate traffic patterns using one or more rule-based approaches. The simulated traffic patterns may be used to generate a road layer level mapping of one or more segments of the SD map data that can be converted into HD map data.