VSLAM Point Cloud Alignment with Open Street Map for 6D Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current self-driving car localization methods relying on expensive RTK-GPS and LiDAR sensors are computationally intensive, leading to detection and localization delays, necessitating a more cost-effective and efficient solution for 6D mapping and localization.
Innovation Solution
The proposed solution involves a fusion of low-cost sensor data using visual simultaneous localization and mapping (VSLAM) techniques, synchronizing GPS timestamps with camera poses, transforming data into Earth-centered, Earth-fixed coordinates, and utilizing Kalman filters to synthesize high-frequency location data, thereby reconstructing point clouds and camera poses for efficient mapping and localization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If expensive RTK-GPS and LiDAR sensors are used for localization, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent replaces expensive, complex sensors (RTK-GPS and LiDAR) with cheaper alternatives (standard GPS and monocular camera). The system uses multiple inexpensive sensor components that individually have lower precision but collectively achieve the required localization accuracy through fusion and VSLAM processing.
Solution Approach 2:
The patent makes the low-cost sensors perform multiple functions. The monocular camera not only captures images for visual processing but also contributes to 3D point cloud reconstruction and pose estimation. The GPS provides both location data and timestamp synchronization, serving multiple purposes in the localization pipeline.
2Productivity
If RTK-GPS and LiDAR data are fused for high-frequency localization, then productivity is improved, but use of energy and computational resources increase
Solution Approach 1:
The patent replaces computationally intensive sensor fusion (RTK-GPS + LiDAR) with a lighter computational approach using standard GPS and monocular camera data. The VSLAM algorithm processes visual data more efficiently than full LiDAR point cloud processing, reducing computational energy requirements while maintaining high localization frequency.
Solution Approach 2:
The patent extracts and removes the most computationally demanding components (RTK-GPS and LiDAR) from the sensor fusion system. By eliminating these heavy computational loads and retaining only essential localization functions through cheaper sensors and optimized algorithms, the system reduces energy consumption while preserving productivity.
3Measurement precision
If GPS and LiDAR data are utilized for 6D mapping, then measurement precision is improved, but loss of time occurs due to computational intensity
Solution Approach 1:
The patent replaces time-consuming LiDAR processing with faster visual processing using a monocular camera. The VSLAM algorithm processes 2D images more quickly than LiDAR processes 3D point clouds, reducing processing delays. The system achieves comparable 6D mapping precision through efficient visual feature extraction and pose estimation.
Solution Approach 2:
The patent substitutes the mechanical/optical LiDAR scanning system with a digital visual processing system. Instead of using physical laser ranging and complex point cloud operations, the system uses camera image processing and computational photography techniques to achieve 6D pose estimation, significantly reducing processing time while maintaining precision.
Data Source
AI summary
A method of mapping and localization is disclosed that includes, reconstructing a point cloud and a camera pose based on VSLAM, synchronizing the camera pose and a GPS timestamp at a first set of GPS coordinate points and transforming the first set of GPS coordinate points corresponding to the GPS timestamp into a first set of ECEF coordinate points. The method also includes determining a translation and a rotation between the camera pose and the first set of ECEF coordinate points, transforming the point cloud and the camera pose into a second set of ECEF coordinates based on the translation and the rotation and transforming the point cloud and the camera pose into a second set of GPS coordinate points. The method further includes constructing and storing a key-frame image, a key-frame timestamp and a key-frame GPS based on the second set of GPS coordinate points.