3D Point Cloud Matching for Precise Driverless Vehicle Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies for controlling the movement of mobile objects, such as vehicles, through autonomous or remote control, face challenges in accurately matching three-dimensional point cloud data to enhance the precision of vehicle positioning and orientation.
Innovation Solution
A control device that utilizes measured and reference point cloud data for accurate matching, incorporating vehicle-specific information like model codes and appearance features, and adjusts point cloud density based on vehicle characteristics and distance, to generate precise control commands for driverless driving.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If three-dimensional point cloud data matching is performed without considering vehicle-specific appearance features, then the matching process is simple, but the positioning and orientation accuracy deteriorates
Solution Approach 1:
The patent segments the point cloud data processing by creating separate processing paths: one for acquiring vehicle-specific appearance feature information and another for performing density-adjusted matching. This segmentation allows the system to incorporate detailed vehicle characteristics without overwhelming the matching algorithm, thereby improving positioning accuracy while managing complexity through structured data organization.
Solution Approach 2:
The system performs preliminary acquisition and processing of vehicle appearance feature information (model codes, body shapes, colors) before the actual point cloud matching occurs. This preliminary action prepares vehicle-specific reference data that can be quickly referenced during matching, enhancing accuracy without adding computational burden during the critical matching phase.
2Measurement precision
If uniform point cloud density is used for all vehicles, then the data processing is consistent, but the matching accuracy for different vehicle types and distances deteriorates
Solution Approach 1:
The patent implements local quality by adjusting point cloud density based on specific vehicle characteristics (model code, body shape, color) and distance from the imaging device. Different regions of the point cloud data are processed with different density levels tailored to the specific vehicle being matched, allowing optimal detail representation for each vehicle type and distance combination while maintaining overall system versatility.
Solution Approach 2:
The system dynamically changes the density parameter of point cloud data based on vehicle appearance features and distance measurements. By modifying this key parameter according to specific vehicle characteristics, the system achieves superior matching accuracy for diverse vehicle types and ranges while maintaining a unified processing framework that adapts to different scenarios.
3Measurement precision
If detailed vehicle appearance information is acquired and used in matching, then the matching accuracy improves, but the information processing time increases
Solution Approach 1:
The system performs preliminary acquisition and storage of vehicle appearance feature information (model codes, body shapes, colors) in a database before matching operations. This preliminary action allows the matching algorithm to quickly retrieve and compare pre-processed vehicle characteristics without performing complex feature extraction during real-time matching, thereby improving accuracy while minimizing processing time delays.
Data Source
AI summary
A control device used to generate a control command to cause a mobile object to move by driverless driving includes a matching section configured to perform matching using measured point cloud data and reference point cloud data in order to generate the control command, and a mobile object information acquisition section configured to acquire mobile object information by which an appearance feature of the mobile object is identified. The matching section is configured to perform the matching according to the mobile object information. The measured point cloud data is three-dimensional point cloud data on the mobile object acquired by using a distance measurement device. The reference point cloud data is date representing three-dimensional point cloud data for reference.