LIDAR Odometry With Dynamic Surfel Maps for Accurate Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in accurately localizing themselves within their environment due to occasional sensor input errors, especially over high-mileage operations, which can affect the reliability and precision of their navigation and motion planning.
Innovation Solution
A LIDAR odometry system generates a real-time local environment map using LIDAR observations, aligning them with a local environment map to determine a transform between a vehicle frame and a keyframe, minimizing distances between LIDAR points and surfels to provide accurate pose estimates, and periodically updating the map to account for changing surroundings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional localization methods using sensor data are used, then the system can operate with simple processing, but accuracy deteriorates over high-mileage operations due to sensor input errors
Solution Approach 1:
The patent creates a virtual copy of the environment by generating a 3D surfel map from LIDAR observations. Instead of relying on potentially erroneous sensor inputs for localization, the system uses this replicated environmental model to determine pose, thereby decoupling localization accuracy from direct sensor reliability.
Solution Approach 2:
The surfel map serves as an intermediary between the vehicle's LIDAR observations and the localization calculation. By transforming raw sensor data into this intermediate environmental representation, the system can perform more reliable pose estimation that is less sensitive to individual sensor errors.
2Adaptability or versatility
If a static map is used for localization, then processing is simple, but the system cannot account for changing surroundings
Solution Approach 1:
The patent transforms the static map into a dynamic structure by continuously updating the surfel map with new LIDAR observations as the vehicle moves. This allows the environmental model to adapt to changing surroundings while maintaining a manageable representation through real-time updates rather than complete remapping.
Solution Approach 2:
The environment is segmented into discrete surfels (surface elements) that can be individually updated and managed. This segmentation allows the system to handle complex environmental changes by updating only relevant portions of the map rather than processing the entire environment at once.
3Speed
If real-time pose estimation is performed, then navigation responsiveness is improved, but computational resources are consumed
Solution Approach 1:
The system performs preliminary actions by pre-processing LIDAR observations to build and maintain the surfel map structure before final pose estimation is needed. This preparation allows the actual pose calculation to be more efficient, as it only needs to align the map with current sensor data rather than process raw data from scratch.
Data Source
AI summary
A localization system can obtain a LIDAR observation oriented relative to a vehicle frame, the vehicle frame oriented with respect to a pose of a vehicle; access a local environment map descriptive of the environment of the vehicle, wherein the local environment map is oriented relative to a keyframe at a given time, the local environment map including a plurality of surfels and generated in real-time during a current operational instance of the vehicle based on one or more prior LIDAR observations captured during the current operational instance of the vehicle; determine a transform between the vehicle frame and the keyframe by aligning the LIDAR observation to the local environment map based on a similarity between the LIDAR observation and the local environment map; and determine an updated pose of the vehicle based on the transform and the pose of the vehicle in the vehicle frame at the given.