Steering Reliability Mapping From Curvature Discrepancies
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles rely on geometrical map data that does not assess safety and reliability of road situations, leading to potential discomfort and safety issues when discrepancies arise between mapped and experienced curvatures, particularly in unstable or dangerous conditions.
Innovation Solution
A system that determines differences between experienced and geometry-based curvatures by aggregating curvature samples from multiple vehicles and processing them with sensor and crowd-sourced data to identify areas of low reliability, allowing for adjustments in vehicle operation, such as reverting to manual control in problematic areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If autonomous vehicles use geometry-based map data for navigation, then the vehicle can operate autonomously along planned paths, but the vehicle may encounter safety and reliability issues when the actual road conditions differ from the mapped geometry
Solution Approach 1:
The system collects curvature samples from multiple vehicles traversing the same road segments and uses this feedback to create a steering reliability map that compares expected geometry-based curvature with actual experienced curvature. This feedback mechanism allows the system to identify discrepancies between mapped road geometry and real-world driving conditions, enabling safety assessments and predictive maneuvers.
Solution Approach 2:
The system pre-processes curvature samples from multiple vehicles and generates a steering reliability map before autonomous vehicles encounter potentially dangerous situations. By calculating reliability metrics and identifying problematic road segments in advance, the system can prepare predictive maneuvers and alert drivers before entering unstable or dangerous road situations.
2Measurement precision
If the system aggregates curvature samples from multiple vehicles to improve road condition assessment, then the accuracy of road condition identification improves, but the system complexity and data processing requirements increase
Solution Approach 1:
The system uses a multi-vehicle curvature sampling approach where data from multiple vehicles serves multiple purposes: improving measurement precision of road conditions, creating the steering reliability map, and enabling predictive maneuvers. This universal use of aggregated data justifies the increased system complexity by providing multiple benefits from the same data collection infrastructure.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An approach is provided for comparing experienced curvatures with geometry-based curvatures to identify road environments. The approach involves causing, at least in part, an aggregation of a plurality of curvature samples collected from one or more vehicles traversing one or more travel segments. The approach also involves processing and/or facilitating a processing of the curvature samples to determine at least one experienced curvature for the one or more travel segments. The approach further involves determining at least one geometry-based curvature for the one or more travel segments. The approach also involves determining one or more differences between at least one experienced curvature and the at least one geometry-based curvature.