Vehicle Bounding Box Estimation from LiDAR Convex Hull Tangents
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle extent estimation methods using LiDAR measurements are computationally complex and slow due to non-linear optimization and grid search approaches, which involve re-sorting and re-calculating distances for each bounding box orientation.
Innovation Solution
Employing convex hulls to identify bounding box orientations where measurement points switch sides, determining local minimum distances, and selecting the orientation with a global minimum to fit the measurement points, thereby optimizing candidate orientations and reducing computational resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If non-linear optimization methods and grid search approach are used for bounding box orientation estimation, then measurement precision is improved, but device complexity and computation time increase
Solution Approach 1:
The patent segments the continuous orientation search space into discrete angular intervals. Instead of performing non-linear optimization over the entire continuous space, the method divides the orientation range into multiple discrete intervals and performs simplified calculations for each interval, reducing the overall computational complexity while maintaining estimation accuracy.
Solution Approach 2:
The patent performs preliminary sorting of measurement points and pre-calculation of geometric properties before the actual orientation estimation. By pre-processing the data and organizing it in a structured manner, the subsequent orientation search and distance calculations become significantly faster, reducing the computational burden of the main estimation algorithm.
2Measurement precision
If non-linear optimization methods are used for bounding box orientation estimation, then measurement precision is improved, but processing speed decreases
Solution Approach 1:
The patent replaces the expensive non-linear optimization computation with a series of simpler, computationally cheaper operations. Instead of iteratively optimizing the orientation parameters using complex mathematical methods, the method uses discrete interval sampling and basic geometric calculations that are much faster to compute, sacrificing minimal accuracy for significant speed improvement.
Solution Approach 2:
The patent performs preliminary sorting of measurement points and pre-calculation of geometric properties before the actual orientation estimation. By pre-processing the data and organizing it in a structured manner, the subsequent orientation search and distance calculations become significantly faster, reducing the computational burden of the main estimation algorithm.
3Measurement precision
If all measurement points are re-sorted and re-calculated for every orientation, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The patent performs preliminary sorting of measurement points and pre-calculation of geometric properties once, before the orientation estimation process. These pre-computed results are then reused across multiple orientation intervals, eliminating the need to re-sort and re-calculate for every orientation check. This significantly reduces computation time while preserving the accuracy of distance calculations.
Solution Approach 2:
The patent merges the sorting and distance calculation operations into a single pre-processing step. Instead of performing separate sorting and calculation operations for each orientation, the method combines these operations once upfront, and then uses the results efficiently across all orientation intervals, reducing redundant computations.
Data Source
AI summary
Provided are methods for efficient vehicle extent estimation, which can include bounding box generation. Some methods described include determining bounding boxes surrounding detected point clusters according to tangents to convex hulls of the point clusters, and minimizing continuous functions of distances between points and bounding box sides. Accordingly, best-fit bounding boxes are determined more efficiently and quickly, as well as more accurately. Systems and computer program products are also provided.