Scanned Lidar Velocity Determination via Feature Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in quickly determining the velocities of objects in their environment due to limited response times and processing delays from complex segmentation and object tracking algorithms, which can hinder timely maneuvering and operational parameter adjustments.
Innovation Solution
An imaging system that generates two sets of features at different time intervals using overlapping scan lines to compute candidate velocities, allowing for rapid determination of object velocities without relying on full frame processing or complex segmentation, thereby improving situational awareness and reducing computational load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full frame processing and complex segmentation algorithms are used to determine object velocities, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The patent extracts only the essential information needed for velocity determination by using overlapping scan lines to identify corresponding features between two time intervals. Instead of processing the entire frame through complex segmentation algorithms, the system extracts discrete feature points from the overlapping region and computes velocities directly from these extracted features, thereby reducing processing time while maintaining velocity measurement precision.
2Measurement precision
If complex segmentation and object tracking algorithms are used, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent replaces complex mechanical-like processing systems (full frame segmentation and object tracking algorithms) with a simpler computational approach. By using overlapping scan lines to directly compute velocity from feature correspondences between two time intervals, the system substitutes the complex multi-step segmentation and tracking pipeline with a more direct calculation method that achieves velocity determination with reduced algorithmic complexity.
3Measurement precision
If full frame processing is performed to calculate velocities, then measurement precision is improved, but productivity decreases
Solution Approach 1:
The patent applies partial action by processing only the overlapping portion of scan lines from two time intervals rather than performing full frame processing. This partial processing approach focuses computational resources on the essential overlapping region where velocity information can be extracted, thereby maintaining velocity determination accuracy while significantly improving response speed and productivity.
Data Source
AI summary
A scanning imaging sensor is configured to sense an environment through which a vehicle is moving. A method for determining one or velocities associated with objects in the environment includes generating features from the first set of scan lines and the second set of scan lines, the two sets corresponding to two instances in time. The method further includes generating a collection of candidate velocities based on feature locations and time differences, the features selected pairwise with one from the first set and another from the second set. Furthermore, the method includes analyzing the distribution of candidate velocities, for example, by identifying one or more modes from the collection of the candidate velocities.