Coverage-Guided Road Selection and Prioritization for Efficient Testing in Autonomous Driving Systems

TL;DR

The paper tackles the inefficiency of exhaustive ADAS testing by proposing a coverage-guided, behavior-aware road testing framework that preserves geometric and dynamic diversity while reducing redundancy. It combines curvature-based road segmentation, DTW-based geometric similarity, and dynamic driving metrics to cluster road segments, select representative tests, and rank them using a multi-criteria score that balances geometry, dynamics, and historical failures. Empirical results on OpenCat/Udacity demonstrate substantial test-suite reductions (up to 96%) with strong retention of failure exposure (APFD up to $0.97$) and dramatic improvements in early fault detection (up to $\sim$95× vs random baselines), though model transferability is limited and largely model-specific in its failure patterns. The approach offers a practical path to cost-effective, scalable ADAS regression testing by prioritizing tests that maximize behavioral diversity and likelihood of exposing faults, enabling faster safety validation and more efficient use of computational resources.

Abstract

Autonomous Driving Assistance Systems (ADAS) rely on extensive testing to ensure safety and reliability, yet road scenario datasets often contain redundant cases that slow down the testing process without improving fault detection. To address this issue, we present a novel test prioritization framework that reduces redundancy while preserving geometric and behavioral diversity. Road scenarios are clustered based on geometric and dynamic features of the ADAS driving behavior, from which representative cases are selected to guarantee coverage. Roads are finally prioritized based on geometric complexity, driving difficulty, and historical failures, ensuring that the most critical and challenging tests are executed first. We evaluate our framework on the OPENCAT dataset and the Udacity self-driving car simulator using two ADAS models. On average, our approach achieves an 89% reduction in test suite size while retaining an average of 79% of failed road scenarios. The prioritization strategy improves early failure detection by up to 95x compared to random baselines.

Figures (5)

• Figure 1: Coverage-based selection and prioritization framework.
• Figure 2: Illustrative example of the proposed framework.
• Figure 3: Geometric-only vs Hybrid.
• Figure 4: Cross-model test failure consistency analysis for the selected tests.
• Figure 5: Cross-Model Failure Distribution (Dave-2/Chauffeur).