Topological Graph Smoothing for Autonomous Navigation in Semi-Structured Environments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous navigation in semi-structured environments, such as parking lots, is challenging due to the lack of predefined lane-network graphs, requiring the development of methods to estimate and construct obstacle-free driving paths from sensor data.
Innovation Solution
A method involving the creation of a grid-based map using sensor data to form a graph representing the driving path, applying discrete heuristic transformations, locally and globally smoothing the graph, and imposing a lane network to optimize the driving path while avoiding obstacles, utilizing techniques like Generalized Voronoi Diagrams and statistical processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the vehicle follows the lane-network graph in semi-structured environments, then safety is improved by staying on predefined paths, but time and distance efficiency deteriorate due to inability to take optimal shortcuts
Solution Approach 1:
The system dynamically adjusts the lane-network graph based on real-time sensor data and environmental conditions. The graph is not static but can be modified to include temporary shortcuts or alternative paths when obstacles are detected, allowing the vehicle to maintain safety while improving time efficiency through adaptive path recalculation
Solution Approach 2:
The patent introduces an intermediary layer between the predefined lane-network graph and the vehicle's actual path. This intermediary consists of dynamically generated guidance paths that bridge the gap between the structured graph and optimal shortcuts through semi-structured areas, allowing the vehicle to safely navigate efficiency-improving paths while maintaining reference to the underlying safe graph structure
2Measurement precision
If the system creates detailed obstacle-free diagrams and smooths graphs extensively, then path accuracy is improved, but computational overhead increases
Solution Approach 1:
The obstacle-free diagram creation process is segmented into multiple stages: initial graph generation from sensor data, discrete heuristic transformations, local smoothing of individual edges, and global smoothing of the entire graph. This segmentation allows computational resources to be distributed across different processing stages, maintaining high path accuracy while managing computational overhead through progressive refinement rather than attempting all processing simultaneously
Solution Approach 2:
The system applies smoothing operations selectively rather than uniformly across the entire graph. Local smoothing is applied only to edges that require refinement based on specific geometric criteria, and global smoothing is applied only when necessary to resolve larger-scale path inconsistencies. This partial action approach maintains path accuracy where needed while avoiding unnecessary computational expenditure in areas where the graph is already sufficiently accurate
Data Source
AI summary
A method of creating an obstacle-free diagram using topological sensor data to form a graph corresponding to a driving path, transforming the graph using discrete heuristics, locally smoothing a plurality of edges of the graph after the transforming, and globally smoothing the graph after the locally smoothing. Transforming includes deleting an edge of the graph, merging two intersections of the graph into a single intersection, and collapsing a plurality of edges forming a loop into a single edge. Locally smoothing includes smoothing a lane segment of the graph by minimizing a sum of a function of smoothness and a function of distance. Globally smoothing includes defining an intersection potential and minimizing a sum of the function of smoothness, the function of distance, a function of intersections and a function of direction.