LIDAR Ego-Motion Correction for Accurate Object Velocity Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in navigating due to the vast amount of data they need to process, analyze, and store, including visual information, GPS data, and sensor data, which can limit their navigation capabilities.
Innovation Solution
The use of cameras and LIDAR systems to provide autonomous vehicle navigation features, where processors analyze images and point cloud data to determine navigational responses, such as object detection and velocity calculation, and utilize sparse maps for navigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional mapping technology is used for autonomous vehicle navigation, then navigation capability is provided, but the volume of data needed to store and update the map becomes excessively large
Solution Approach 1:
The patent extracts only the essential navigational elements from traditional maps, creating sparse maps that contain only critical information needed for autonomous navigation. This selective extraction reduces data volume while maintaining navigation capability by focusing on key features rather than storing complete environmental data.
Solution Approach 2:
Instead of storing complete map data and querying it for navigation, the system inverts the approach by using LIDAR point cloud data to dynamically generate and update sparse maps only where needed. This inversion allows the system to maintain navigation capability with minimal data storage requirements.
2Measurement precision
If LIDAR systems perform multiple scans to detect moving objects, then object detection accuracy improves, but the processing time and computational complexity increase
Solution Approach 1:
The system performs preliminary processing by storing raw LIDAR point cloud data from multiple scans without immediate full processing. This allows the system to accumulate data over time and process it efficiently later, reducing real-time processing time while maintaining the ability to detect moving objects with high accuracy through comparison of accumulated scans.
3Measurement precision
If complete point cloud data from multiple LIDAR scans is processed, then object velocity calculation accuracy improves, but the computational load and processing time increase
Solution Approach 1:
The patent extracts only the necessary information from complete point cloud data for velocity calculation. Instead of processing all points in multiple scans, the system identifies and processes only relevant features and changes between scans, significantly reducing computational load while maintaining velocity calculation accuracy through selective analysis of point cloud differences.
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 autonomous vehicles to efficiently process and interpret environmental data, improve navigation accuracy, and reduce the storage and processing demands associated with traditional mapping technologies.
Implementation Method 1
receive, from a LIDAR system associated with the host vehicle and based on a first LIDAR scan of a field of view of the LIDAR system, a first point cloud including a first representation of at least a portion of an object
Data Source
AI summary
A navigation system for a host vehicle may include a processor programmed to determine at least one indicator of ego motion of the host vehicle. A processor may be also programmed to receive, from a LIDAR system, a first point cloud including a first representation of at least a portion of an object and a second point cloud including a second representation of the at least a portion of the object. The processor may further be programmed to determine a velocity of the object based on the at least one indicator of ego motion of the host vehicle, and based on a comparison of the first point cloud, including the first representation of the at least a portion of the object, and the second point cloud, including the second representation of the at least a portion of the object.