Hierarchical Geometric Occupancy Mapping for Collision Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current occupancy mapping technologies, such as voxel grids, spatial hash tables, and octrees, face challenges in efficiently representing and updating dynamic environments, requiring significant time or memory resources, and often limit the representation of obstacles to uniform cells, making them inefficient for complex or cluttered spaces.
Innovation Solution
The use of geometric entities with hierarchical relationships organized in a tree structure, where geometric entities at different levels represent occupied spaces, allowing for efficient collision detection and obstacle representation using a hash function to extract semantic information like occupancy and velocity, and employing a learning-based method to generate these entities from point-cloud data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If voxel grids are used for occupancy mapping, then the representation is simple and effective for fast look-up and insertion, but the memory usage grows logarithmically with map size, limiting them to small spaces
Solution Approach 1:
The patent segments the occupancy space into multiple levels of hierarchical representation, where each level divides space into smaller regions. This segmentation allows the system to represent large spaces efficiently by only storing occupied regions at appropriate levels, rather than uniformly gridting the entire space, thus reducing memory usage while maintaining fast access speeds.
Solution Approach 2:
The patent introduces a hierarchical dimension to the traditional voxel grid structure. Instead of a single-level 3D grid, the system creates multiple hierarchical levels where each level provides a coarser or finer representation of space. This dimensional transformation allows efficient representation of both small and large spaces by navigating through different hierarchical levels, reducing overall memory requirements.
2Speed
If spatial hash tables are used for sparse representation, then dynamic expansion is achieved with constant time insertion and lookup, but maintaining sparsity in dynamic environments is expensive
Solution Approach 1:
The patent implements a dynamic hierarchical structure that automatically adapts to changes in the environment. When objects are added or removed, the system dynamically updates only the affected hierarchical regions rather than maintaining the entire structure. This dynamic approach preserves constant-time insertion and lookup while reducing the overhead of maintaining sparsity by focusing computational effort only on changed areas.
Solution Approach 2:
The patent applies different levels of detail and update frequency to different regions of the occupancy space based on their importance and dynamism. Frequently changing or important regions receive more detailed hierarchical representation and more frequent updates, while static or less important regions use coarser representation. This local quality differentiation reduces the overall computational cost of maintaining sparsity while preserving fast access where needed.
3Quantity of substance
If octrees are used to provide complete hierarchical data structure, then memory usage is reduced for static environments, but fine-grained voxelization is still needed in cluttered environments and updates are expensive
Solution Approach 1:
The patent segments the octree structure into variable-resolution regions based on occupancy characteristics. Instead of uniformly applying fine-grained voxelization throughout the entire space, the system identifies cluttered or important regions and applies detailed segmentation only where necessary, while using coarser representation in open spaces. This selective segmentation reduces memory usage and simplifies hierarchy maintenance by avoiding unnecessary fine-grained divisions.
Solution Approach 2:
The patent dynamically changes the resolution parameter of the octree structure based on local occupancy density and importance. In cluttered environments, the system increases the resolution parameter to provide fine-grained representation, while in open areas, it decreases the resolution to reduce complexity. This parameter adaptation allows the system to maintain an efficient hierarchical structure that balances memory usage, update cost, and representation accuracy without requiring complete fine-grained voxelization throughout.
Data Source
AI summary
A system can map occupancy of objects. The system may generate an occupancy representation of an object based on a point cloud of the object. The system may generate first geometric entities based on the point cloud. Each first geometric entity contains one or more points in the point cloud. The system may also generate one or more second geometric entities, each of which contains one or more first geometric entities. The occupancy representation of the object includes the one or more second geometric entities and the plurality of first geometric entities. The occupancy representation may have a hierarchical structure where the first geometric entities may be on a lower level than the one or more second geometric entities. The system can also detect collision of the object with another object by using the occupancy representation of the object and an occupancy representation of the other object.