Cooperative LiDAR Relative Pose Estimation Using Odometry and ICP
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining relative pose between vehicles using LIDAR sensors are computationally burdensome, require HD maps that may be outdated or unavailable, and rely on GPS data that is not always accessible, making real-time point cloud stitching challenging.
Innovation Solution
A cooperative perception system that computes a base relative pose using an iterative closest point (ICP) algorithm at a low frequency (1 Hz) and combines it with odometry-based temporal relative poses computed at a higher frequency (10 Hz) to generate a combined point cloud, allowing real-time stitching without continuous high-frequency ICP execution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If iterative closest point (ICP) algorithm is executed continuously at high frequency to compute relative pose, then measurement precision of relative pose estimation is improved, but computational burden increases and real-time processing becomes difficult
Solution Approach 1:
The patent segments the relative pose computation into two distinct components: a base relative pose computed once at low frequency using ICP algorithm, and temporal relative poses computed at high frequency using odometry data. This segmentation allows the computationally intensive ICP to run infrequently while maintaining real-time performance through the lightweight odometry-based temporal poses.
Solution Approach 2:
The patent performs preliminary computation of the base relative pose at a lower frequency before computing temporal relative poses. This preliminary action establishes a reference frame that simplifies subsequent real-time computations, allowing the system to achieve high-frequency output without continuously executing the full ICP algorithm.
2Measurement precision
If HD map is used to determine absolute pose for relative pose estimation, then measurement precision is improved, but device complexity and data requirements increase
Solution Approach 1:
The patent extracts and eliminates the dependency on HD maps from the relative pose estimation system. By using odometry data from vehicle sensors, the system can compute temporal relative poses without requiring external HD map references, thereby reducing system complexity and data requirements while maintaining precision.
3Measurement precision
If GPS data is used to improve relative pose estimation, then measurement precision is improved, but reliability decreases in environments where GPS is unavailable
Solution Approach 1:
The patent implements a self-service system that uses odometry data generated by the vehicle's own sensors (LIDAR, IMU) to compute temporal relative poses. This eliminates dependency on external GPS signals, allowing the system to maintain reliable operation in GPS-denied environments such as tunnels and urban canyons by using the vehicle's intrinsic sensing capabilities.
Data Source
AI summary
System, methods, and other embodiments described herein relate to cooperative perception. In one embodiment, a method includes computing, at a first timestep, a base relative pose between an ego vehicle and a remote vehicle based upon respective point clouds of the ego vehicle and the remote vehicle. The method includes computing, at a second timestep, a relative pose between the ego vehicle and the remote vehicle based upon the base relative pose, a first temporal relative pose of the ego vehicle, and a second temporal relative pose received from the remote vehicle. The method includes generating a combined point cloud based upon a first point cloud of the ego vehicle, a second point cloud received from the remote vehicle, and the relative pose.