Autonomous Vehicle Path Planning via Adaptive Candidate Point Density
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for planning pass paths in unmanned driving technologies lead to unnecessary calculation of impassable paths, increasing the calculation amount and decreasing planning speed, and result in poor coupling between the planned path and the driving strategy due to lack of consideration for driving strategies.
Innovation Solution
Acquire current road environment information and set pass candidate points based on both the environment and driving strategy, determining an optimal passable path to reduce unnecessary calculations and improve path planning speed while enhancing coupling with driving strategies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If pass candidate points are set homogeneously in all areas in front of the vehicle, then the path planning covers all possible areas, but the calculation amount increases and planning speed decreases due to unnecessary calculation of impassable paths
Solution Approach 1:
The patent applies local quality by setting pass candidate points with different densities in different regions. Specifically, the area in front of the vehicle is divided into multiple regions, and pass candidate points are set sparsely in regions far from the vehicle and densely in regions close to the vehicle. This resolves the contradiction by maintaining comprehensive coverage (reliability) while reducing unnecessary calculations in distant areas (productivity).
Solution Approach 2:
The patent segments the road environment into multiple regions based on distance from the vehicle and obstacle positions. By dividing the planning space into different zones with different point densities, the system avoids homogeneous sampling and reduces calculation overhead in less critical areas while maintaining detail where needed.
2Device complexity
If pass candidate points are set without considering driving strategy, then the path planning is simple, but the coupling between planned path and driving strategy is poor
Solution Approach 1:
The patent applies preliminary action by pre-establishing the relationship between driving strategies and pass candidate point settings. The system determines the current driving strategy first, then uses this strategy to guide the setting of pass candidate points. This ensures that the path planning is inherently aligned with the driving strategy without requiring complex post-processing or re-planning.
Data Source
AI summary
The present disclosure provides a method, an apparatus, a device and a readable storage medium for planning a pass path. The method includes: acquiring current road environment information in front of a vehicle; setting pass candidate points according to the current road environment information and a driving strategy; and determining a current optimal passable path according to the pass candidate points. Since the pass candidate points are set according to the current road environment information and the driving strategy, and the current optimal passable path is determined according to the pass candidate points, thereby effectively decreasing the calculation amount of the impassable path calculation in prior art due to direct setting of pass candidate points in all areas in front of an unmanned vehicle.