Connected and Automated Vehicle (CAV) Testing Scenario Design and Implementation Using Naturalistic Driving Data and Augmented Reality
archive: archived pipeline: cataloged verified
Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)
Summary
This paper addresses the challenge of efficiently and accurately testing Connected and Automated Vehicles (CAVs) by proposing a unified framework for Test Scenario Library Generation (TSLG). Existing testing standards, such as Federal Motor Vehicle Safety Standards, are inadequate for CAVs because they assume human drivers and fail to evaluate vehicle "intelligence." Public road testing is inefficient due to the rarity of critical scenarios, while simulation lacks physical realism. The authors aim to solve the TSLG problem by generating a library of critical scenarios that allows for comprehensive evaluation of CAV performance metrics—including safety, functionality, mobility, and rider comfort—with fewer tests than traditional methods. The methodology utilizes Naturalistic Driving Data (NDD) and an Augmented Reality (AR) testing platform. The framework defines scenario criticality as a product of maneuver challenge (probability of an event of interest, such as an accident, given a scenario) and exposure frequency (probability of the scenario occurring in real traffic). To identify critical scenarios, the authors employ a surrogate model calibrated on human driving data to estimate maneuver challenge, combined with NDD for exposure frequency. A novel search method uses an auxiliary objective function and a "seed-fill" algorithm to locate local critical scenarios in the decision variable space, avoiding the inefficiencies of random sampling. These scenarios form a library from which tests are sampled using an $\epsilon$-greedy policy to balance exploitation of known critical scenarios and exploration of new ones. Testing is conducted in a closed facility using an AR platform that synchronizes real CAV movements with simulated background vehicles. The study validates the framework through two case studies: a cut-in scenario for safety evaluation and a highway exit scenario for functionality evaluation. In the cut-in case, the library generation and evaluation process demonstrated that the proposed method achieves the same accuracy as NDD-based evaluation but accelerates the process by 103 times. The results confirmed that the method provides unbiased index estimation of performance metrics with significantly fewer required tests. The highway exit case further illustrated the framework's applicability to functionality metrics, defining task difficulty to quantify maneuver challenges for task completion. The significance of this work lies in providing a generic, efficient, and accurate method for CAV validation that overcomes the limitations of existing approaches, which often rely on specific CAV models or focus solely on safety. By integrating NDD and AR, the framework enables the evaluation of CAV intelligence across multiple performance metrics in a controlled environment. The theoretical justification for the criticality definition and the demonstrated acceleration in testing efficiency offer a scalable solution for the validation of automated driving systems, supporting the development and deployment of safer and more capable CAVs.
Key finding
The proposed framework for generating test scenario libraries accelerates CAV evaluation by 103 times compared to naturalistic driving data methods while maintaining the same accuracy.
Methodology
mixed_methods
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | rosap | — | — | 2 | 2026-05-23 |
| archive | success | — | — | — | 1 | 2026-05-23 |
| extract | success | cached | — | — | 2 | 2026-06-10 |
| clean | success | — | — | — | 1 | 2026-06-01 |
| chunk | success | — | — | — | 1 | 2026-06-01 |
| embed | success | — | — | — | 1 | 2026-06-02 |
| enrich | success | — | — | — | 1 | 2026-05-23 |
| promote | success | — | — | — | 1 | 2026-05-23 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 3 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 19 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Methodological Resource: tool software, validation psychometrics, dataset resource