Automated Driving Model Testing With Progressive Obstacle Scenarios
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing number of possible scenarios in automated driving systems makes rigorous testing burdensome, limiting the number of vehicles that can be considered in realistic scenarios, especially with multiple obstacles, which poses a challenge in ensuring safety and efficiency in navigation.
Innovation Solution
A method and system that prune scenarios by initially testing single-obstacle scenarios, identifying those that could result in collisions, and iteratively adding more obstacles, focusing only on scenarios where additional obstacles could cause collisions, thereby reducing the number of scenarios to be tested and accelerating the training and refinement of automated driving models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all possible scenarios with multiple obstacles are tested, then safety and reliability of automated driving models are improved, but the complexity and time required for testing increases significantly
Solution Approach 1:
The testing process is segmented into multiple phases: first testing single-obstacle scenarios, then progressively adding obstacles only to scenarios that previously resulted in collisions. This segmentation divides the overwhelming testing task into manageable stages, reducing overall testing complexity while maintaining safety verification.
Solution Approach 2:
Single-obstacle scenarios are tested and evaluated before adding additional obstacles. Scenarios that do not result in collisions are identified and excluded from further multi-obstacle testing. This preliminary action filters out safe scenarios early, preventing unnecessary complexity in subsequent testing phases.
2Reliability
If all possible scenarios with multiple obstacles are tested, then reliability is improved, but the time required for testing increases significantly
Solution Approach 1:
Single-obstacle scenarios are tested first to establish a baseline safety evaluation. Scenarios that pass this preliminary test (do not result in collisions) are excluded from time-consuming multi-obstacle testing. This preliminary action significantly reduces total testing time while maintaining reliability verification for critical scenarios.
Solution Approach 2:
The testing timeline is segmented into discrete phases where only necessary scenarios progress to the next phase. By separating single-obstacle testing from multi-obstacle testing and only advancing problematic scenarios, the overall testing duration is reduced without compromising safety verification.
3Adaptability or versatility
If the number of obstacles in scenarios is increased, then realism and comprehensiveness of testing are improved, but the number of scenarios to be tested increases dramatically
Solution Approach 1:
Single-obstacle scenarios serve as a preliminary filter to identify which scenarios require further multi-obstacle testing. By evaluating scenarios with fewer obstacles first, the system determines which specific scenarios need additional obstacle complexity, thereby reducing the total number of comprehensive scenarios that must be tested.
Solution Approach 2:
Scenario complexity is segmented and added progressively rather than all at once. Scenarios start with one obstacle and additional obstacles are added only to scenarios that demonstrate collision potential. This segmentation maintains scenario comprehensiveness for critical cases while dramatically reducing the total quantity of scenarios requiring full multi-obstacle testing.
Data Source
AI summary
Methods and systems for deploying an automated driving model include testing an automated driving model with a set of scenarios that include a single obstacle. The set of scenarios is pruned to include only those scenarios that do not result in a collision, but that may result in a collision with the addition of another obstacle. New scenarios are added to the set of scenarios to include scenarios that have an additional obstacle to form an updated set of scenarios. The automated driving model is tested with the updated set of scenarios. The automated driving model is deployed to a vehicle to control operation of one or more systems in the vehicle.