Traffic Scenario Similarity Graphs for Efficient Autonomous Vehicle Testing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for testing autonomous vehicle functions are time-consuming and costly, requiring extensive real-world testing that is impractical and inefficient, and existing simulation methods struggle with duplicate and similar scenarios, leading to excessive testing times.
Innovation Solution
A computer-implemented method using a similarity classification module and machine learning techniques to classify scenario similarities, generating an efficient execution sequence for testing autonomous vehicle systems by identifying and prioritizing unique scenarios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If extensive real-world testing is conducted to verify autonomous vehicle functions, then testing coverage and reliability are improved, but time consumption and costs increase significantly
Solution Approach 1:
The patent creates virtual copies of real-world traffic scenarios through simulation environments. Instead of physically testing every possible driving situation, the system generates synthetic scenario data that replicates real traffic conditions, allowing comprehensive testing coverage without the time and cost constraints of actual road testing.
Solution Approach 2:
The system performs preliminary classification and selection of test scenarios before actual execution. By pre-processing scenarios to identify duplicates and similarities, the system prepares an optimized test sequence that ensures comprehensive coverage while minimizing redundant testing, thus reducing overall time consumption.
2Reliability
If all scenarios from databases with thousands of scenarios are tested, then scenario coverage is improved, but testing time becomes excessive due to duplicates and similar scenarios
Solution Approach 1:
The system extracts and removes duplicate and highly similar scenarios from the test database before execution. By identifying scenarios with similarity above a threshold and retaining only representative instances, the system eliminates redundant testing while preserving comprehensive scenario coverage, thus improving testing efficiency.
Solution Approach 2:
The system automatically classifies and optimizes the test scenario sequence without manual intervention. The automated similarity detection and sequence optimization processes enable the system to self-manage the testing workflow, reducing the need for human curation while maintaining high testing efficiency and coverage.
3Productivity
If massive parallelization is used to test all scenarios, then testing speed is improved, but the overall process remains too time-consuming due to the large number of scenarios
Solution Approach 1:
The system performs partial classification and optimization of scenarios rather than processing every single scenario. By applying similarity detection and sequence optimization to a representative subset or using threshold-based filtering, the system achieves sufficient test coverage with reduced computational overhead, making the overall process feasible even with massive parallelization.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Computer-implemented method for providing a test process for traffic scenarios to be tested, having the steps of: training and using a similarity classification module for the purpose of classifying the similarity between two scenarios based on a subset of a parameter set of the particular scenario; generating a graph representation comprising a multiplicity of scenarios, wherein nodes each represent the subset of a parameter set of a scenario and edges between two nodes are weighted with a similarity classification value which indicates the similarity between the two scenarios of the respective nodes; receiving a user input by means of a user interface and/or an automatic system input, wherein the input comprises a required value for covering the scenarios and/or a required number of scenarios to be tested, with the result that a selection and an execution sequence of the scenarios are determined.