Dynamic Reference Line Sampling Density for Autonomous Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous driving vehicle systems face computational inefficiencies due to fixed reference line sampling point densities, which can lead to increased computation time for trajectory planning and decision-making, especially when dealing with complex road environments.
Innovation Solution
A computer-implemented method that dynamically adjusts the sampling point density of reference lines for autonomous driving vehicles by using a nonlinear algorithm to select reference points closer to the vehicle at a higher density and farther away at a lower density, reducing the number of points required for accurate trajectory planning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If fixed reference line sampling point density is used, then reference line generation is simple, but computation time for trajectory planning increases
Solution Approach 1:
The patent applies dynamics by transitioning from fixed reference line sampling point density to dynamic sampling point density that adapts to the vehicle's current state. The sampling point density is adjusted based on vehicle speed, curvature of the reference line, and distance from the vehicle, allowing the system to optimize computation time while maintaining trajectory planning accuracy.
Solution Approach 2:
The patent implements parameter changes by modifying the sampling point density parameter based on multiple factors including vehicle speed, reference line curvature, and distance from the vehicle. The system dynamically adjusts the sampling interval parameter to reduce computation time in straight sections while maintaining higher density in curved sections where precision is critical.
2Measurement precision
If high sampling point density is used throughout the reference line, then trajectory accuracy is improved, but computational load increases
Solution Approach 1:
The patent applies local quality by implementing non-uniform sampling point density along the reference line. High sampling point density is concentrated in regions close to the vehicle and in curved sections where trajectory accuracy is critical, while lower density is used in distant straight sections, optimizing the balance between accuracy and computational load.
Solution Approach 2:
The patent implements partial action by applying high sampling point density only where necessary for accurate trajectory planning (near the vehicle and in curved sections) rather than uniformly across the entire reference line. This selective approach reduces overall computational load while maintaining sufficient accuracy for safe autonomous operation.
3Ease of manufacture
If uniform sampling distance is used, then reference line generation is straightforward, but computation time increases for long reference lines
Solution Approach 1:
The patent transitions from static uniform sampling distance to dynamic sampling distance that varies along the reference line. The sampling distance is adjusted based on the vehicle's distance from each reference point, with closer points having smaller sampling distances and farther points having larger sampling distances, optimizing processing efficiency.
4Measurement precision
If dense sampling points are used far from the vehicle, then future trajectory precision is improved, but current computation time increases
Solution Approach 1:
The patent applies local quality by concentrating sampling points in regions that are locally relevant to the vehicle's current trajectory planning needs. High density sampling is applied near the vehicle where immediate trajectory control is critical, while distant regions use lower density, reducing current computation time while maintaining sufficient precision for near-term planning.
Solution Approach 2:
The patent implements partial action by applying dense sampling only to the portion of the reference line that is currently relevant for trajectory planning (within a certain distance from the vehicle) rather than uniformly across the entire reference line, reducing computation time while maintaining adequate precision for safe operation.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
According to some embodiments, a system receives a first set of reference points based on a map and a route information, the first plurality of reference points representing a reference line in which the ADV is to follow. The system selects a second set of reference points along the reference line, including iteratively performing, selecting a current reference point from the first set of reference points, determining a sampling distance along the first set of reference points based on the currently selected reference point using a nonlinear algorithm, and selecting a next reference point based on the determined sampling distance such that a density of the second set of reference points closer to the ADV is higher than a density of the selected reference points farther away from the ADV. The system plans a trajectory for the ADV using the second set of reference points to control the ADV.