Autonomous Vehicle Point Cloud Map Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Constructing high-definition (HD) maps with precise detail and accuracy for autonomous vehicles is challenging due to the need for data-intensive sensors and dynamic object tracking, which affects navigation and localization.
Innovation Solution
A method and system for constructing and controlling autonomous vehicles using a two-step pose difference computation process, aligning three-dimensional point cloud map segments, and controlling the vehicle based on these alignments, incorporating coarse-granularity and fine-granularity pose differences, and image-plane re-projection error minimization techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data-intensive sensors like LiDARs and Radars are used to construct HD maps, then measurement precision and manufacturing precision are improved, but device complexity and loss of energy increase
Solution Approach 1:
The patent segments the map construction process into multiple stages: offline HD map generation, online map segment retrieval, and real-time alignment. By dividing the complex task of constructing and using HD maps into manageable segments, the system reduces computational burden and energy consumption while maintaining high precision through specialized processing at each stage
Solution Approach 2:
The patent performs preliminary actions by pre-generating offline HD maps and pre-retrieving relevant map segments before real-time navigation. This allows the vehicle to work with pre-processed, organized data during critical real-time operations, reducing on-board computational complexity and energy requirements while preserving measurement precision
2Measurement precision
If HD maps contain exact locations of all intersections, road signs, and dynamic objects, then measurement precision is improved, but loss of information and device complexity increase
Solution Approach 1:
The patent applies local quality by retrieving and processing only the specific map segment relevant to the vehicle's current location and task, rather than managing complete HD maps containing all possible details. This selective approach maintains measurement precision for relevant areas while reducing information management complexity by excluding unnecessary data
Solution Approach 2:
The patent extracts only the necessary map segment from the complete offline HD map based on the vehicle's current context and navigation needs. This extraction process removes unnecessary information while preserving the precision required for the specific task, thereby reducing data management complexity without sacrificing measurement accuracy
3Manufacturing precision
If a two-step pose difference computation process is used for aligning map segments, then manufacturing precision is improved, but loss of time and device complexity increase
Solution Approach 1:
The patent segments the pose difference computation into two distinct steps: coarse alignment and fine alignment. This segmentation allows the system to first quickly establish a rough alignment (coarse step) and then refine it (fine step), achieving high manufacturing precision while managing computation time efficiently by addressing different precision requirements at different stages
Data Source
AI summary
Systems and method are provided for controlling an autonomous vehicle. In one embodiment, a method includes: receiving sensor data from a sensor of the vehicle; determining a three dimensional point cloud map segment from the sensor data; determining a vehicle pose associated with the three-dimensional point cloud map segment; determining a pose difference based on the vehicle pose, another vehicle pose, and a two-step process, wherein the two-step process includes computing a coarse-granularity pose difference, and computing a fine-granularity pose difference; aligning the three dimensional point cloud map segment with another three dimensional point cloud map segment associated with the other vehicle pose based on the pose difference; and controlling the vehicle based on the aligned three dimensional point cloud map segments.