Road Segment Templates for Autonomous Driving Certification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The regulatory process for certifying high-level autonomous driving has not converged, limiting the use of high-level autonomous vehicles to restricted settings, and there is a need to extend the whitelist of road segments deemed safe for such operations by identifying candidate road segments with similar properties to existing whitelisted segments.
Innovation Solution
The system uses nonparametric clustering and geometric hashing techniques to establish road segment templates based on features like geometry, topology, visibility, and traffic patterns, and scores candidate road segments for contiguity and continuity to reconstruct safe road segments for autonomous operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If high-level autonomous driving is permitted only in restricted settings with a whitelist of approved road segments, then safety and regulatory control are improved, but the extent of autonomous driving deployment and utility are limited
Solution Approach 1:
The system performs preliminary analysis and scoring of candidate road segments before adding them to the whitelist. By pre-evaluating segments using automated scoring based on geometric hashing and template matching, the system prepares a ranked list of candidates that can be efficiently reviewed and approved, thus expanding deployment while maintaining safety controls.
Solution Approach 2:
The system creates templates from existing whitelisted road segments and uses these templates to identify and evaluate similar candidate segments. By copying the proven safety characteristics of approved segments into templates, the system can efficiently certify new segments that match these templates, thereby expanding deployment extent while maintaining consistent safety standards.
2Reliability
If manual review and certification of each road segment is performed, then regulatory control and safety are improved, but the time and resources required for certification increase
Solution Approach 1:
The system enables candidate road segments to self-certify by automatically comparing themselves against established templates and generating their own safety scores. This automated self-evaluation reduces the burden on regulators, who only need to review pre-scored candidates rather than manually assess every segment from scratch, thus maintaining regulatory control while reducing certification time.
Solution Approach 2:
The system transforms the certification process by changing from qualitative manual assessment to quantitative automated scoring. By converting road segment characteristics into measurable parameters that can be automatically compared against templates, the system maintains rigorous regulatory control through objective criteria while dramatically reducing the time required for certification.
3Adaptability or versatility
If the whitelist of approved road segments is expanded to increase autonomous driving utility, then deployment flexibility is improved, but the risk of including unsafe segments increases
Solution Approach 1:
The system applies different levels of scrutiny to different candidate segments based on their similarity to established templates. Highly similar segments receive automated approval with standard safety assurance, while segments with partial matches or unique characteristics receive enhanced review. This differentiated approach allows expansion of the whitelist while maintaining appropriate safety assurance for each segment.
Solution Approach 2:
The system implements a feedback mechanism where certified road segments are continuously monitored and their performance data feeds back into the template refinement process. This creates a closed-loop system where the whitelist expansion is guided by real-world performance feedback, ensuring that safety assurance is maintained and improved over time as the system learns from actual autonomous driving operations.
Data Source
AI summary
Methods, devices and apparatuses pertaining to identifying the candidate road segments for autonomous operations are described. A method may involve obtaining data of a first plurality of road segments that are permitted for one or more autonomous operations. The method may further include generating a road segment template based on the data of the first plurality of road segments, and reconstructing a second plurality of road segments based on the road segment template to obtain a plurality of reconstructed road segments.