LiDAR-to-HD Map Transition Mapping for Precise Vehicle Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
LiDAR-based vehicular localization systems face challenges in accurately determining vehicle position relative to a mapped environment due to distortions caused by aligning LiDAR point clouds with pre-existing high-definition maps, which can reduce localization accuracy and deform object shapes, affecting object detection.
Innovation Solution
A method involving a transition map that defines six degrees of freedom transformations between LiDAR maps and high-definition maps, allowing for precise alignment without distorting the LiDAR point cloud, enabling accurate vehicle localization by applying these transformations to determine corresponding locations on the high-definition map.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If LiDAR point cloud is distorted to fit pre-existing map, then localization alignment is improved, but point cloud crispness and object shape accuracy deteriorate
Solution Approach 1:
The patent introduces a transition map as an intermediary data structure that stores pre-computed transformations between LiDAR coordinate systems and HD map coordinate systems. This transition map acts as a mediator that enables accurate localization alignment without requiring real-time distortion of the LiDAR point cloud, thus preserving point cloud crispness while achieving proper alignment with the HD map.
Solution Approach 2:
The patent pre-computes and stores transformations in the transition map during map creation or update phases, rather than computing them in real-time during localization. This preliminary action allows the system to have transformation data ready before localization is needed, enabling fast and accurate alignment without distorting the actual LiDAR point cloud data during operation.
2Measurement precision
If LiDAR point cloud is distorted to match HD map, then coordinate alignment is improved, but object shape recognition deteriorates
Solution Approach 1:
The transition map serves as an intermediary that handles coordinate transformation mathematically without physically distorting the LiDAR point cloud. By applying transformation matrices stored in the transition map, the system achieves coordinate alignment between LiDAR and HD map while keeping the original point cloud data intact, thus preserving object shape accuracy for recognition algorithms.
Solution Approach 2:
Instead of distorting the LiDAR point cloud to fit the HD map, the patent inverts the approach by transforming HD map coordinates to LiDAR coordinates using the transition map. This inversion allows the system to query HD map data in the LiDAR coordinate system, achieving alignment without deforming the actual sensor data and maintaining object shape fidelity.
3Measurement precision
If LiDAR point cloud is distorted for map fitting, then localization accuracy is improved, but mapping system optimizations are negated
Solution Approach 1:
The transition map acts as an intermediary layer that decouples the LiDAR mapping system from the HD map coordinate system. By introducing this intermediate transformation layer, the patent allows the LiDAR mapping system to maintain its original optimizations and produce high-quality point clouds without distortion, while still achieving accurate localization through the transition map's coordinate transformations.
Solution Approach 2:
The patent segments the localization function into two independent parts: (1) LiDAR point cloud acquisition and mapping using optimized LiDAR mapping algorithms, and (2) coordinate system transformation using the transition map. This segmentation allows each component to operate independently with its own optimizations intact, preventing the negation of mapping system performance while achieving accurate localization.
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 maintains the crispness and accuracy of LiDAR maps, improving localization precision by aligning LiDAR data with high-definition maps without distorting the original data, thus enhancing object recognition and vehicle positioning in complex environments.
Implementation Method 1
LiDAR is one type of sensor that may be used for vehicular localization that emits light beams (e.g., laser beams) and may detect light beam reflections to determine the distance of objects in the environment
Data Source
AI summary
The present disclosure generally relates to methods and systems for determining a position of a vehicle relative to a surrounding environment. A vehicle may obtain a LiDAR map of the surrounding environment at a vehicle location via processing data from a LiDAR device mounted on the vehicle. The vehicle may access a transition map based on the vehicle location to determine a transformation between the LiDAR map and a stored high-definition map. The transition map may define a six degrees of freedom transformation for each of a plurality of locations. The vehicle may apply the transformation to an element of the LiDAR map to determine a corresponding location of the element on the high-definition map. In some aspects, the vehicle may perform an autonomous driving operation based on a location of the vehicle with respect to the element of the high-definition map.