Train here, drive there: ROS based end-to-end autonomous-driving pipeline validation in CARLA simulator using the NHTSA typology
DOI: 10.1007/s11042-021-11681-7
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 challenge of validating fully autonomous driving (AD) architectures in complex urban environments, where physical testing is prohibitively expensive and time-consuming. Motivated by the high rate of human-error-related accidents and the need for robust decision-making systems, the authors propose a simulation-based validation pipeline. The study focuses on integrating a ROS-based autonomous driving architecture with the CARLA simulator, utilizing the NHTSA crash typology to define specific, challenging driving scenarios. The primary goal is to evaluate the system’s behavioral decision-making module, which relies on Hierarchical Interpreted Binary Petri Nets (HIBPN), as a preliminary step before deployment on a real-world electric vehicle prototype. The methodology employs a modular architecture built on the Robot Operating System (ROS) and Docker containers to ensure isolation, flexibility, and portability. The system comprises four layers: hardware drivers, control (including localization, mapping, and reactive control via Pure Pursuit and Beam Curvature Method algorithms), executive (decision-making), and interface. The decision-making layer uses HIBPNs to manage concurrent behaviors and transitions, avoiding the state explosion problem associated with finite-state machines. The authors integrated this architecture with CARLA, an open-source simulator based on Unreal Engine, by adapting the CARLA ROS Bridge to match the sensor configurations of their real vehicle. Perception was simulated using fused LiDAR and camera data, with semantic segmentation provided by CARLA’s pseudo-sensors to identify obstacles and drivable areas. The validation process tested the architecture against four NHTSA-inspired use cases: Pedestrian Crossing, Stop, Adaptive Cruise Control (ACC), and Unexpected Pedestrian. The authors presented both qualitative results, such as simulation videos, and quantitative analyses, including linear velocity profiles and trajectory data split by HIBPN states. Additionally, they analyzed temporal graphs associated with Vulnerable Road Users (VRU) to assess the system's responsiveness and safety in dynamic scenarios. The results demonstrated that the HIBPN-based decision-making module could effectively handle these complex interactions, maintaining safe distances and appropriate speeds while navigating intersections and reacting to sudden obstacles. The significance of this work lies in its demonstration of a robust, end-to-end validation framework that bridges the gap between simulation and real-world deployment. By leveraging hyper-realistic simulation and standardized crash typologies, the authors provide a scalable approach to testing AD systems that mitigates the risks and costs associated with physical trials. The study highlights the importance of using modular, containerized architectures and advanced decision-making models like HIBPNs to handle the uncertainty and complexity of urban driving. This pipeline serves as a solid baseline for future research, offering a reproducible method for validating autonomous navigation systems before they are implemented in physical prototypes.
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 | 2026-06-26 |
| extract | success | cached | — | — | 2 | 2026-06-26 |
| clean | success | clean | — | — | 1 | 2026-06-20 |
| chunk | success | chunk | — | — | 1 | 2026-06-20 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-20 |
| promote | success | — | — | — | 1 | 2026-06-19 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-26 |
| tag | success | vector_similarity | — | — | 6 | 2026-06-20 |
| 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).
- Methodological Resource: tool software
- Theoretical Contribution: computational model