Semantic Map Obstacle Localization Using Waypoint Screening
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional methods for locating obstacles in semantic maps for autonomous driving have low accuracy, which affects navigation strategies and overall driving performance.
Innovation Solution
A method and apparatus that acquire obstacle coordinates and travel velocity, select an initial region in the semantic map, and apply a screening rule to identify a target waypoint based on travel velocity, road segment directions, distances, and connectivity, using machine learning for point cloud data identification and KD tree storage for efficient querying.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional coordinate transformation methods are used to locate obstacles in semantic maps, then the implementation is simple, but the locating accuracy is low
Solution Approach 1:
The patent segments the obstacle localization process into multiple stages: initial region selection based on obstacle coordinates, waypoint screening using multiple criteria (distance, velocity consistency, road segment direction, connectivity), and iterative refinement. This segmentation transforms a single complex coordinate transformation into a series of manageable filtering and selection steps, improving accuracy while maintaining implementability
Solution Approach 2:
The patent performs preliminary actions by first selecting an initial region around the obstacle coordinates and pre-screening waypoints based on distance thresholds before the final localization. This preliminary filtering reduces the search space and eliminates obviously incorrect waypoints early in the process, improving both accuracy and computational efficiency
2Measurement precision
If multiple screening criteria are applied to waypoints, then the locating accuracy is improved, but the computational complexity increases
Solution Approach 1:
The patent applies preliminary distance-based filtering to eliminate waypoints that are obviously too far from the obstacle coordinates before applying more computationally intensive screening criteria such as velocity consistency and road segment direction analysis. This preliminary action reduces the number of waypoints requiring complex evaluation, balancing accuracy with computational power
Solution Approach 2:
The patent implements a multi-level screening process where not all waypoints undergo all screening criteria. Instead, waypoints are progressively filtered through multiple passes of increasingly stringent criteria, applying full computational power only to the most promising candidates rather than uniformly processing all waypoints
Data Source
AI summary
Provided are a method and apparatus for locating an obstacle in a semantic map, a computer device, and a storage medium. The method includes: acquiring obstacle coordinates of an obstacle and a travel velocity of the obstacle; selecting an initial region in the semantic map according to a preset selection radius and according to the obstacle coordinates; and performing screening on a plurality of waypoints in the initial region by using a preset screening rule so as to obtain a target waypoint matching a position of the obstacle in the semantic map, and using the target waypoint as the position of the obstacle in the semantic map.