Navigation Occupancy Data Using Voxels for Precise Vehicle Positioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current navigation data structures struggle to efficiently model and store the geometry of road furniture, such as guardrails and street lamps, in highly automated driving systems, which is crucial for precise vehicle positioning and obstacle detection, while maintaining compact storage and supporting updates.
Innovation Solution
The implementation of a navigation data source that uses space-filling curves to linearly order sub-regions of a spatial region, allowing for efficient spatial queries and compact storage, utilizing voxels to represent occupancy information and interval data to optimize storage and memory usage, compatible with existing navigation data standards.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If detailed 3D modeling of road furniture is implemented using traditional data structures, then positioning accuracy and obstacle detection improve, but data storage size and memory consumption increase significantly
Solution Approach 1:
The spatial region is divided into discrete sub-regions (voxels) that can be independently represented and stored. Each voxel indicates occupancy status rather than storing complete geometric models, segmenting the detailed 3D information into manageable binary units that reduce overall data volume while preserving spatial structure for accurate positioning.
Solution Approach 2:
The patent transitions from storing traditional 2D map data to a 3D voxel-based representation. By adding the vertical dimension and using space-filling curves to map 3D coordinates to 1D array indices, the system achieves efficient compact storage of spatial occupancy information while maintaining the ability to query and process three-dimensional spatial relationships for accurate vehicle positioning.
2Measurement precision
If comprehensive spatial data of roadside objects is stored for precise vehicle positioning, then positioning accuracy improves, but data processing time and computational complexity increase
Solution Approach 1:
The patent extracts only the essential occupancy information (present/absent) from complete 3D models of roadside objects. By storing merely the presence/absence indicators in each voxel rather than full geometric descriptions, the system reduces computational complexity for spatial queries while retaining sufficient information for accurate vehicle positioning and obstacle detection.
Solution Approach 2:
Different regions of space are represented with uniform voxel resolution appropriate for the specific application requirements. The system applies this consistent level of detail throughout the spatial region, optimizing the balance between positioning precision and processing efficiency by using the same simplified representation strategy across all areas where vehicle positioning is needed.
3Adaptability or versatility
If traditional data structures are used for spatial information, then compatibility with existing systems is maintained, but spatial query efficiency and memory usage deteriorate
Solution Approach 1:
The voxel-based spatial occupancy grid serves multiple functions: it enables efficient spatial queries through 1D array indexing via space-filling curves, provides compact storage by using binary occupancy indicators, and maintains compatibility with existing navigation data standards through standardized data interfaces. This multi-functional approach allows the same data structure to deliver both query efficiency and system compatibility.
Solution Approach 2:
The patent introduces space-filling curves as an intermediary mathematical transformation that maps 3D spatial coordinates to 1D array indices. This intermediary mechanism enables efficient indexing and querying of the voxel data while maintaining the underlying 3D spatial relationships, bridging the gap between compact 1D storage and efficient 3D spatial queries without requiring complex data structures.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention inter alia relates to a navigation data source, said navigation data source (100) comprising: an object data set (101), said object data set (101) comprising object data (103), wherein said object data (103) indicating a spatial vacancy and/or occupancy of sub-regions (501) of a spatial region (500) by one or more structural objects in said spatial region (500), and wherein said object data (103) references said sub-regions (501) based on a linear order of said sub-regions (501) in said spatial region (500). The object data (103) may comprise interval information (104) about an at least partially occupied interval (520), in particular a lower interval border and/or an upper interval border, wherein said at least partially occupied interval (520) indicates a group of, according to said linear order, one or more successional sub-regions (501), at least a part of which are spatially occupied, and wherein said object data (103) comprises an occupancy sequence (105) indicating a spatial vacancy and/or occupancy of said successional sub-regions (501) of said at least partially occupied interval (520).