LiDAR Heading Angle Determination for Close-Range Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing LiDAR-based systems face challenges in accurately determining the heading angle of objects, especially when only a portion of the object is within the detection range, leading to incorrect shape and movement detection, particularly for vehicles cutting in or overtaking from a close distance.
Innovation Solution
A method using a weighted sum of candidate straight lines determined from LiDAR points, where weights are proportional to the length of the candidate straight lines, integrated to improve the accuracy of heading angle determination, and a LiDAR-based object detection device with a controller for performing object detection and tracking within a detection range.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the existing LiDAR-based object detection method is used, then the detection range is sufficient for remote objects, but the heading angle determination becomes inaccurate for close-distance objects
Solution Approach 1:
The patent applies different processing strategies based on the distance to the object. For close-distance objects where only partial point data is available, it uses a simplified bounding box method that relies on extreme points rather than full object contour analysis. This local adaptation of the detection method to the quality of available data resolves the contradiction between measurement precision and information loss.
Solution Approach 2:
The patent changes the detection parameters based on object distance. For remote objects with complete point data, it uses comprehensive contour analysis for accurate heading angle determination. For close objects with incomplete data, it switches to using only extreme points (minimum and maximum coordinates) to define a bounding box, thereby adapting the measurement approach to the available information quality.
2Measurement precision
If the same heading angle determination method is applied to all objects, then the process is simple, but the detection accuracy varies for objects at different positions
Solution Approach 1:
The patent introduces dynamic adaptability by making the detection method flexible based on object position and data completeness. Rather than a static single-method approach, the system dynamically selects between comprehensive contour analysis for remote objects and extreme-point bounding box method for close objects, thereby achieving both precision and versatility.
Solution Approach 2:
The patent creates a universal detection framework that can handle both remote and close-distance objects effectively. By incorporating two different determination methods within one system and automatically selecting the appropriate method based on object distance and point data completeness, the system achieves multi-functionality that adapts to various object positions and data conditions.
3Area of stationary object
If LiDAR points are obtained only for a limited portion of the object, then the detection range is maintained, but the shape and movement detection becomes incorrect
Solution Approach 1:
The patent extracts only the essential information needed for accurate detection when full object data is unavailable. For close-distance objects with limited point data, it extracts the extreme points (minimum and maximum coordinates in each dimension) to define a bounding box, rather than attempting to reconstruct the complete object shape. This extraction of critical information maintains detection accuracy despite limited data coverage.
Solution Approach 2:
The patent applies partial action by using only the necessary subset of available LiDAR points for detection. Rather than requiring complete object coverage, it successfully detects objects using only the extreme points that define the bounding box boundaries, thereby achieving accurate detection with partial data while maintaining the detection range.
Data Source
AI summary
A method for detecting based on LiDAR an object around a vehicle which includes a Light Detection and Ranging (LiDAR) sensor includes determining a first object from a point cloud within a detecting range, and determining a heading angle of the first object by a first method, wherein the first method including determining candidate straight lines for the heading angle of the first object, and determining the heading angle of the first object from angles of the candidate straight lines.