Voxel-Based Multiresolution Ground Plane Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in accurately determining the ground surface due to discarding excessive lidar data when invalidating large three-dimensional planes, leading to misidentification of environment features and potential transition to manual driving modes.
Innovation Solution
A ground surface manager system that associates lidar data with a voxel space, determines multiresolution planes, validates them for accuracy, and connects valid planes to create a precise representation of the ground surface, ensuring accurate data retention and reduced misclassification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If large three-dimensional planes are invalidated to improve ground detection accuracy, then measurement precision is improved, but loss of information increases due to discarding excessive lidar data
Solution Approach 1:
The patent divides the ground detection process into multiple resolution levels (coarse and fine), where different plane sizes are evaluated at different scales. This segmentation allows the system to retain useful information from various resolution levels rather than discarding all data from invalidated planes, thus improving ground detection accuracy while minimizing information loss.
Solution Approach 2:
The patent applies different validation criteria and plane size thresholds at different resolution levels and spatial locations. Instead of uniformly invalidating all large planes, the system selectively validates planes based on local ground characteristics and resolution requirements, preserving valuable lidar data while maintaining detection accuracy in critical regions.
2Measurement precision
If multiple resolution levels are processed to improve ground surface representation, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent implements a hierarchical processing architecture that segments the multiresolution analysis into distinct coarse and fine resolution levels. Each level processes planes of appropriate sizes independently, reducing the computational complexity compared to processing all planes at all resolutions. This segmented approach maintains high measurement precision while managing device complexity through structured organization.
Solution Approach 2:
The patent dynamically adjusts the plane validation process based on resolution level, selecting appropriate plane size thresholds and validation criteria for each level. This dynamic adaptation allows the system to optimize processing complexity at each stage while maintaining overall measurement precision, avoiding the need for fixed complex processing of all possible plane configurations.
Data Source
AI summary
Techniques for determining ground surface voxels based on multiresolution planes are discussed herein. Such multiresolution planes may be generated, validated, and used to connect voxels associated with the ground. In some examples, a vehicle may use capture or otherwise receive lidar data while traversing within a driving environment. The vehicle may associate the lidar data with a voxel space. Voxels can be associated with a cell of a multiresolution grid and the vehicle may determine multiresolution planes associated with the multiresolution space. For example, the vehicle may determine a first plane associated with a first resolution level, in addition to determining a plurality of planes associated with a second resolution level. The vehicle may determine the validity of the plane(s) at the various resolution levels. Based on the validity of the planes, the vehicle may connect the valid planes that are associated with the ground surface.