Virtual Testing in Automated Driving Systems Certification. A Longitudinal Dynamics Validation Example
DOI: 10.1109/access.2022.3171180
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This paper addresses the critical need for validating virtual testing environments for Automated Driving Systems (ADS) certification. While virtual testing offers safety and efficiency advantages over physical testing, it lacks standardized validation methodologies to ensure simulation-generated evidence is reliable. The authors focus on quantifying the fidelity of a Vehicle-Hardware-in-the-Loop (VeHIL) setup compared to real-world experiments, specifically examining longitudinal dynamics. The study aims to provide a quantitative, end-to-end validation procedure that can be generalized to other toolchains, moving beyond qualitative comparisons or ADS-specific validation. The methodology involved a comparative analysis between a physical proving ground (PG) test and a VeHIL replication. A robotized SMART ForTwo equipped with a camera-based ADS was tested on a proving ground in Italy, collecting data via GNSS units for longitudinal maneuvers: free-flow driving, car-following, and stop-and-go. The PG scenarios were reconstructed and replicated in the VeHIL environment, which used a dyno-chassis and a wide-screen monitor to stimulate the vehicle’s camera with photorealistic simulations from Vires VTD. Because the VeHIL setup lacked steering capabilities, lateral dynamics were excluded, limiting the scope to longitudinal Key Performance Indicators (KPIs): velocity and acceleration. The authors employed several statistical metrics to assess fidelity, including Normalized Root Mean Square Error (NRMSE), Pearson correlation, coefficient of determination ($R^2$), and the Sprague-Geers metric for time-series comparison. Additionally, two-sample T-tests and Kolmogorov-Smirnov (KS) tests were applied to aggregated data to evaluate statistical equivalence. The results demonstrated an overall good match between real-world and simulated data. In the free-flow scenario, the VeHIL showed higher repeatability with narrower confidence intervals than the physical tests. Statistical testing on aggregated data yielded high p-values for both T-tests and KS-tests, supporting the null hypothesis that the virtual and physical samples derive from the same distribution, thus validating the environment. However, without data aggregation, the KS-test questioned the validity, highlighting the sensitivity of statistical methods to data conditioning. The study also noted limitations in the vehicle model, which relied on coast-down curves rather than full dynamic models, restricting accuracy during transient effects and mild decelerations. The significance of this work lies in its provision of a rigorous, quantitative framework for validating virtual testing environments, a prerequisite for their adoption in ADS certification. By demonstrating that state-of-the-art VeHIL setups can achieve high fidelity for longitudinal dynamics, the paper supports the development of "virtual proving grounds." The authors conclude that while the results are encouraging, open issues remain in defining a complete validation framework, particularly regarding the extrapolation of models to scenarios different from those used in validation. The methodology presented offers a scalable approach for assessing simulation toolchains, contributing to the broader goal of safe and efficient ADS certification.
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.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | OpenAlex-citations | — | — | 1 | 2026-06-19 |
| archive | success | unpaywall | — | — | 2 | 2026-06-25 |
| extract | success | cached | — | — | 2 | 2026-06-26 |
| clean | success | clean | — | — | 1 | 2026-06-19 |
| chunk | success | chunk | — | — | 1 | 2026-06-19 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-19 |
| promote | success | — | — | — | 1 | 2026-06-19 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-26 |
| tag | success | vector_similarity | — | — | 6 | 2026-06-19 |
| verify | success | — | — | — | 1 | 2026-06-26 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; 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).
- Applied Guidance: standards test procedures
- Methodological Resource: validation psychometrics, tool software