Autonomous Driving Trajectory Generation Using SD and HD Map Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous driving technologies face challenges in accurately generating optimal paths using high-definition maps due to large data sizes, and struggle with efficient lane changes and collision avoidance under adverse conditions, particularly when objects obstruct the driving path.
Innovation Solution
A method that integrates navigation route data from standard definition maps with high-definition map lane centerlines to generate autonomous driving trajectories, prioritizing lane changes and avoiding obstacles by utilizing both map types, adjusting paths dynamically based on connectivity and obstacle presence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-definition map data is used to generate autonomous driving paths, then path accuracy is improved, but data processing complexity increases
Solution Approach 1:
The patent segments the high-definition map data into essential geometric elements (lane centerlines, road boundaries, intersection patterns) and processes only these critical components rather than the entire map dataset. This segmentation approach maintains path accuracy while reducing computational complexity by focusing on the most relevant spatial information for autonomous navigation.
Solution Approach 2:
The patent extracts only the necessary path generation elements from the high-definition map data, such as lane centerlines and intersection geometries, while discarding redundant information. This extraction process preserves the accuracy needed for precise path following while significantly reducing the data processing burden on the autonomous driving system.
2Productivity
If lane change operations are delayed until necessary, then fewer lane changes are made, but collision risk increases when objects obstruct the path
Solution Approach 1:
The patent performs preliminary detection of objects that may obstruct the intended driving path and proactively initiates lane change operations before the vehicle reaches the obstruction. By anticipating potential conflicts and executing lane changes in advance, the system maintains driving efficiency while ensuring collision avoidance through timely positional adjustments.
Solution Approach 2:
The patent continuously monitors the driving environment for objects that may block the current or future path, using this feedback information to dynamically adjust lane change timing. The system processes real-time object detection data and determines optimal lane change moments that balance efficient progress with safe avoidance of obstacles.
3Adaptability or versatility
If vision camera data is heavily relied upon for lane recognition, then real-time adaptability is improved, but reliability deteriorates under adverse conditions such as weather disturbances
Solution Approach 1:
The patent introduces pre-generated autonomous driving path data as an intermediary reference that complements vision camera observations. This path data, derived from high-definition maps, provides a reliable baseline trajectory that remains valid under adverse weather conditions, while vision data provides real-time adjustments. The combination maintains both real-time adaptability and reliability by using the map-based path as a stable reference framework.
Data Source
AI summary
A point-to-point autonomous driving path generation and driving control method based on a navigation route and a high definition map, which generates a point-to-point autonomous driving path by utilizing both of a route made with a navigation map (SD map) and a lane centerline of a high definition map (HD map). The point-to-point autonomous driving path generation and driving control method based on the navigation route and the high definition map includes a step of acquiring navigation route data; and a step of generating autonomous driving trajectory data on the basis of lane information of the high definition map matched with the navigation route data.