Autonomous Driving Scenario Retrieval for ODD-Compliant Test Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for validating and assessing autonomous vehicles, particularly those with SAE Level 4 or higher autonomous driving systems, face challenges in efficiently generating and testing various real driving scenarios to ensure safety and operational design domain compliance, as they rely on limited operational design domains and criticality phenomena analysis.
Innovation Solution
A scenario retrieval-based automatic scenario generation method that retrieves and filters scenarios from a database, calculates similarity, converts components to suit target conditions, and concretizes scenarios to create realistic and relevant driving scenarios for testing, ensuring compliance with operational design domains and criticality phenomena.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If scenario-based testing is used to validate autonomous vehicles, then safety and operational design domain compliance are improved, but the complexity of generating and managing diverse driving scenarios increases
Solution Approach 1:
The system pre-stores a large number of driving scenarios in a database before actual testing begins. These scenarios are organized and structured in advance, allowing the testing system to efficiently retrieve and utilize them during validation processes without needing to generate scenarios on-demand, thus reducing operational complexity while maintaining comprehensive safety coverage
Solution Approach 2:
The system creates modified copies of existing stored scenarios by applying transformations to generate new test cases. Instead of manually creating entirely new scenarios, the system replicates and adapts proven scenarios, maintaining consistency and reliability while expanding scenario diversity for thorough validation
2Adaptability or versatility
If real driving situations are made into scenarios for testing, then the realism and relevance of tests are improved, but the time and resources required to create and manage these scenarios increase
Solution Approach 1:
The system captures real driving situations and stores them as standardized scenario templates in the database. These real-world scenarios are then replicated and modified for various testing purposes, maintaining realism while eliminating the need to manually recreate identical situations, thus preserving authenticity without proportional time investment
Solution Approach 2:
The system transforms real driving situations into structured scenarios by adjusting and standardizing parameters such as environmental conditions, vehicle states, and traffic patterns. This parameterization allows real-world complexity to be preserved in a manageable, reusable format that can be efficiently retrieved and modified for different testing needs
3Reliability
If continuous driving is tested in large-scale simulation regions, then the comprehensiveness of validation is improved, but the computational resources and time required increase
Solution Approach 1:
The system pre-generates and stores a comprehensive library of driving scenarios that cover various operational design domains and criticality phenomena before actual continuous driving tests. This pre-prepared scenario database enables extended validation periods to be efficiently utilized without requiring proportional increases in real-time computational resources or manual scenario creation time
Data Source
Figure 1
Figure 2
Figure 3
AI summary
There are provided a scenario similarity retrieval-based automatic scenario generation system and method. According to an embodiment, a scenario retrieval-based automatic scenario generation method includes: retrieving scenarios similar to a query scenario from a scenario DB for an autonomous driving test; filtering only scenarios that meet a selection condition from the retrieved scenarios; and converting components of the filtered scenarios to suit a target condition. Accordingly, a desired scenario may be automatically generated by retrieving a scenario similar to a targeted scenario, converting the retrieved scenario, and concretizing the scenario.