Autonomous Vehicle Localization Using 2D Maps and 3D Point Clouds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle localization methods struggle with accurate 3D localization of autonomous vehicles, especially in environments with limited GPS signal coverage and dynamic objects, requiring significant computing resources and unable to provide full 6DOF representation of the vehicle's location.
Innovation Solution
A method that utilizes a LIDAR scanning system to capture 3D point clouds, converts them into 2D images with vertical distribution and intensity channels, and applies particle filtering to estimate the vehicle's position, combining odometry data and 2D reference maps to determine the vehicle's 3D location and orientation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If object tracking or object detection methods are used to identify dynamic objects, then localization accuracy is improved, but computing resource consumption increases significantly
Solution Approach 1:
The patent extracts only the essential static features (curbs, lane markers, building corners) from the environment for localization, excluding dynamic objects and other irrelevant information. This selective extraction reduces computing resource consumption while maintaining localization accuracy by focusing only on the most reliable and stable features.
Solution Approach 2:
The patent applies different processing strategies to different types of environmental features. Static features are used for primary localization, while dynamic objects are excluded. This localized quality approach optimizes computing resources by applying complex object detection only where necessary and using simpler methods for the majority of features.
2Device complexity
If 2D localization methods are used, then computational complexity is reduced, but the ability to provide full 6DOF representation is lost
Solution Approach 1:
The patent extends traditional 2D localization by incorporating elevation (z-coordinate) to achieve 3D localization. This dimensional extension provides full 6DOF representation (x, y, z, roll, pitch, yaw) while maintaining computational efficiency by building upon established 2D methods rather than implementing completely new 3D algorithms.
3Measurement precision
If manual selection of land markers is used, then adaptability to different environments is reduced, but localization precision is improved
Solution Approach 1:
The patent creates a universal localization system that can automatically identify and use multiple types of features (curbs, lane markers, building corners) across different environments. This multi-functional approach maintains high localization precision while improving adaptability by not relying on manually pre-selected land markers for each specific environment.
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 enables efficient and robust 6DOF localization of autonomous vehicles, reducing computational complexity and improving accuracy, even in environments with dynamic objects and limited GPS coverage.
Implementation Method 1
receiving a three-dimensional (3D) point cloud of an environment surrounding the autonomous vehicle captured by a light detection and ranging (LIDAR) scanning system of the autonomous vehicle
Data Source
AI summary
A method and system for localization of an autonomous vehicle in real-time disclosed. A particle filter process is used to generate an estimate of a pose of the autonomous vehicle based on a three-dimensional point cloud, a two-dimensional (2D) feature map and odometry data. A location of the autonomous vehicle in the environment is generated comprising a final pose of the vehicle determined based on the estimate of the pose of the autonomous vehicle and the odometry data, an elevation of the autonomous vehicle generated based on the final pose of the autonomous vehicle and the 2D feature map, and an orientation of the autonomous vehicle generated based on the odometry data.