LiDAR L-Shaped Contour Extraction for Vehicle Free-Space Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional autonomous parking control technologies fail to accurately recognize surrounding space contours in varying environments, leading to deteriorated space recognition and parking control performance.
Innovation Solution
A method and system that conservatively extracts contours of surrounding objects using LiDAR points by determining corner points, segment parameters, and generating L-shaped contours to output spatial information, including contour information, through clustering and singular value decomposition, allowing for accurate recognition of free-space and parking spaces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional autonomous parking control technology assumes an object protruding toward the vehicle and extracts contour information based on this assumption, then the system can operate with a simplified model, but the contour information becomes inaccurate in environments that differ from the assumed situation
Solution Approach 1:
The patent segments the contour extraction process into multiple stages: first identifying corner points from LiDAR data, then determining segment parameters for each corner, and finally generating L-shaped contours. This segmentation allows the system to handle various object shapes and environments more accurately while maintaining computational efficiency through modular processing.
Solution Approach 2:
The patent changes the parameters used for contour extraction from assuming simple protruding objects to using corner point coordinates, segment parameters, and L-shaped contour definitions. By adjusting these parameters based on actual LiDAR measurements rather than fixed assumptions, the system achieves higher accuracy in diverse environments.
2Productivity
If the system uses inaccurate contour information of objects to recognize surrounding space, then the space recognition process remains simple and fast, but the space recognition performance and parking control performance deteriorate
Solution Approach 1:
The patent performs preliminary clustering of LiDAR points into corner points before contour generation. This preliminary action organizes the raw data into meaningful structures that facilitate faster and more accurate contour extraction, improving both speed and reliability of space recognition.
Solution Approach 2:
The patent replaces conventional contour extraction methods with a new approach based on corner point identification and L-shaped contour generation. This substitution introduces mathematical operations (singular value decomposition) to replace less accurate geometric assumptions, achieving both speed and accuracy.
3Measurement precision
If the system extracts contours conservatively in various situations using corner points and segment parameters, then the contour accuracy improves, but the computational complexity increases
Solution Approach 1:
The patent divides the contour generation algorithm into distinct segments: corner point identification, segment parameter determination, and L-shaped contour construction. Each segment handles a specific aspect of the problem, making the overall complex algorithm more manageable and implementable through modular computation.
4Ease of operation
If the system uses a simple contour extraction method assuming protruding objects, then the algorithm remains simple and fast, but it fails to accurately recognize objects in different environmental situations
Solution Approach 1:
The patent creates a universal contour extraction method that works across various environmental situations. The L-shaped contour model with corner points and segment parameters can represent different object shapes and configurations, making the algorithm versatile for diverse parking environments while maintaining a relatively simple computational framework.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach provides accurate spatial information for improved autonomous parking control by effectively forming and outputting L-shaped contours in various scenarios, enhancing the recognition and control of parking spaces.
Implementation Method 1
clustering LiDAR points obtained by clustering Light Detection and Ranging (LiDAR) points
Data Source
AI summary
According to an embodiment of the present disclosure, a method for recognizing a free-space around a vehicle comprises determining a corner point from cluster points of an object based on a line segment connecting a first point and a second point of the cluster points obtained by clustering Light Detection and Ranging (LiDAR) points, determining a segment parameter according to a distance between the line segment and the corner point based on cluster points located at both sides of the corner point, and generating an L-shaped contour of the object based on the segment parameter to output spatial information including contour information.