Adjacent Map Alignment Using Lane Markers for Smooth Vehicle Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Self-driving vehicles face challenges in smooth traveling control when transitioning between adjacent maps due to inherent errors in map information, leading to variations in self-position estimation and potentially causing sudden deceleration, acceleration, or deviation from the target path.
Innovation Solution
A map generation apparatus that combines first and second map information by recognizing the position of lane markers using point cloud and image information, updating the maps to correct for errors and ensure seamless integration of point cloud and road map data, thereby eliminating variations in self-position recognition and enabling smooth travel control across map boundaries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If map information from multiple adjacent maps is used for self-position estimation, then the coverage area is improved, but the estimation accuracy deteriorates due to inherent errors in each map
Solution Approach 1:
The patent performs preliminary alignment of adjacent maps by recognizing lane markers in advance and adjusting map information before self-position estimation. The processor aligns the coordinate systems of multiple maps by detecting lane markers and calculating transformation parameters, so that when vehicles traverse map boundaries, the pre-aligned maps provide continuous and accurate position estimation without sudden jumps or discontinuities.
2Loss of information
If multiple map information sources are integrated, then the map completeness is improved, but the consistency between maps deteriorates due to inherent errors
Solution Approach 1:
The patent uses lane markers as intermediary reference objects to mediate the integration of multiple map information sources. By detecting lane markers that appear in both adjacent maps and using them as common reference points, the system calculates coordinate transformation parameters that ensure consistent positioning across map boundaries. This intermediary approach allows complete map coverage while maintaining information consistency through the shared lane marker references.
3Area of stationary object
If self-position estimation is performed using adjacent map information, then the navigation coverage is improved, but the travel control smoothness deteriorates due to position estimation variations at map boundaries
Solution Approach 1:
The system performs preliminary alignment of adjacent maps by detecting lane markers and calculating coordinate transformations before vehicles reach map boundaries. This advance preparation ensures that when vehicles cross into adjacent maps, the pre-aligned coordinate systems provide continuous position estimation without sudden jumps, enabling smooth travel control throughout the extended navigation coverage area.
Data Source
AI summary
Map generation apparatus includes processor and memory. Memory stores: first/second map information of first/second map of first/second area adjacent to each other. Processor generates first/second map based on first/second traveling history of first/second vehicle in first/second area; and updates at least one of first/second map information so as to combine first/second maps. First map information includes position information of point cloud recognized based on distance information to surrounding objects acquired by first vehicle. Second map information includes position information of lane marker recognized based on image information acquired by second vehicle. Processor recognizes position of lane marker based on first map information stored in memory; and updates at least one of first/second map information stored in memory so as to combine first/second maps based on recognized position of lane marker and position information of lane marker included in second map information.