Vehicle Control for LiDAR Bounding Box Occlusion Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle control systems face challenges in accurately identifying the direction and type of external objects using LiDAR, particularly in determining whether bounding boxes corresponding to these objects are occluded, which hinders effective autonomous driving and driving assistance.
Innovation Solution
A vehicle control apparatus and method that utilizes a sensor to obtain bounding boxes in a first coordinate system and converts them into polygons in a second coordinate system, determining overlap between these polygons based on vertex analysis and angle values to accurately identify occlusions and generate signals for autonomous driving operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Difficulty of detecting and measuring
If LiDAR is used to identify external objects, then object detection capability is improved, but measurement precision of object direction and occlusion status deteriorates
Solution Approach 1:
The patent segments the bounding box identification process into multiple stages: initial bounding box generation, polygon conversion, overlap detection, and occlusion status determination. This segmentation allows each stage to focus on specific aspects, improving overall measurement precision while maintaining detection capability.
Solution Approach 2:
The patent transforms the problem from 3D bounding box analysis to 2D polygon overlap analysis by projecting bounding boxes onto a 2D plane. This dimensionality change simplifies the computational complexity while maintaining accurate occlusion detection, resolving the contradiction between detection capability and measurement precision.
2Measurement precision
If complex algorithms are used to determine occlusion status, then measurement precision is improved, but computational load increases
Solution Approach 1:
The patent extracts only the essential geometric features (vertices and edges of polygons) needed for occlusion determination, discarding unnecessary 3D spatial information. This extraction reduces computational load while maintaining measurement precision by focusing on the minimal required data.
Solution Approach 2:
The patent creates simplified 2D polygon representations as copies of the original 3D bounding boxes. These polygon copies retain the essential spatial relationships needed for occlusion detection while requiring significantly less computational resources to process.
3Measurement precision
If coordinate system transformation is performed, then object identification accuracy is improved, but processing time increases
Solution Approach 1:
The patent performs coordinate system transformation and polygon generation as preliminary actions before the main occlusion detection process. By preparing the polygon representations in advance, the actual occlusion determination can proceed more quickly, reducing overall processing time while maintaining identification accuracy.
Data Source
AI summary
A method for controlling autonomous driving of a vehicle is introduced. The method may comprise obtaining at least two polygons expressed in a second coordinate system, wherein the at least two polygons correspond to at least two bounding boxes of a plurality of bounding boxes expressed in a first coordinate system, and wherein each of the plurality of bounding boxes is associated with a respective one of a plurality of external objects, determining, based on a plurality of vertices forming the at least two polygons, whether the at least two polygons overlap each other, outputting, based on a ratio at which the at least two polygons overlap each other, at least one of the at least two bounding boxes, and generating, based on the at least one of the at least two bounding boxes outputted, a signal indicating that the at least two polygons overlap each other.