Traffic Scene Similarity: a Graph-based Contrastive Learning Approach
This addresses the problem of efficient scenario-based testing for highly automated vehicles, but it appears incremental as an extension to existing contrastive learning methods.
The paper tackles the challenge of reducing redundant test runs for highly automated driving validation by proposing a graph-based contrastive learning approach to map traffic scenes into an embedding space, where thematically similar clusters are formed to identify similar scenes, though no concrete performance numbers are provided.
Ensuring validation for highly automated driving poses significant obstacles to the widespread adoption of highly automated vehicles. Scenario-based testing offers a potential solution by reducing the homologation effort required for these systems. However, a crucial prerequisite, yet unresolved, is the definition and reduction of the test space to a finite number of scenarios. To tackle this challenge, we propose an extension to a contrastive learning approach utilizing graphs to construct a meaningful embedding space. Our approach demonstrates the continuous mapping of scenes using scene-specific features and the formation of thematically similar clusters based on the resulting embeddings. Based on the found clusters, similar scenes could be identified in the subsequent test process, which can lead to a reduction in redundant test runs.