Autonomous Driving Scenario Prioritization for Test Coverage and Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous vehicle testing methods are inefficient and lack comprehensive scenario coverage, leading to high costs and safety concerns due to the complexity of operation scenarios, with existing methods prioritizing efficiency over coverage.
Innovation Solution
An automatic driving acceleration test method that prioritizes scenario generation based on hazard, exposure frequency, and sensitivity, dividing the parameter space to focus on high-priority scenarios, iteratively updating test priorities, and selecting scenarios for testing to balance efficiency and coverage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all scenarios in the vehicle operation zone are extracted and tested, then test coverage is improved, but test efficiency deteriorates and computing power is wasted
Solution Approach 1:
The vehicle operation zone parameter space is segmented into multiple sub-regions based on hazard levels. High-hazard scenarios are isolated and prioritized for testing, while low-hazard scenarios are excluded or tested with lower priority. This segmentation allows the testing system to focus computational resources on critical scenarios, improving both coverage of dangerous scenarios and overall test efficiency.
Solution Approach 2:
The patent changes the parameter of scenario selection from exhaustive enumeration to priority-based filtering using hazard assessment. By introducing hazard level as a filtering parameter, the system transforms the testing approach from testing all scenarios to testing only high-priority scenarios, thereby improving test efficiency while maintaining coverage of critical safety scenarios.
2Productivity
If test scenarios are prioritized based on hazard assessment, then test efficiency is improved, but scenario coverage may be reduced
Solution Approach 1:
Different coverage strategies are applied to different regions of the parameter space. High-hazard regions receive intensive testing with multiple scenarios and higher priority, while low-hazard regions receive minimal or no testing. This local quality approach ensures comprehensive coverage of critical scenarios while reducing overall testing burden, balancing efficiency and coverage.
Solution Approach 2:
The patent applies partial action by testing only a subset of scenarios - specifically those with high hazard levels - rather than performing exhaustive testing. This partial testing approach is sufficient to ensure safety for critical scenarios while avoiding waste of resources on low-risk scenarios, achieving the right balance between coverage and efficiency.
3Reliability
If the parameter space is divided into zones, then scenario generation coverage is improved, but device complexity increases
Solution Approach 1:
The parameter space is segmented into hazard-based zones that can be independently evaluated and tested. Each zone represents a distinct region with characteristic hazard levels, allowing the testing system to manage complexity by treating each zone as a separate unit with its own testing strategy, rather than managing the entire parameter space as a single complex system.
Data Source
AI summary
The disclosure belongs to the technical field of autonomous vehicle, in particular to an automatic driving acceleration test method considering efficiency and coverage, which includes the following steps. Step 1 is definition of scenario test priority. Step 2 is zone division. Step 3 is search within zones. Step 4 is update of scenario test priorities. Step 5 is iterative test. After selecting the automatic driving function to be tested and setting the parameters of the vehicle operation zone, the scenario generation range is formed. The coverage of the test scenario is improved by dividing the generated range and setting the freedom of early autonomous driving exploration. The efficiency of the test process is improved by continuously improving the probability of generating dangerous scenarios in the test process. Thus, it is ensured that the generated test scenarios take into account both test efficiency and test coverage.