Dynamic driving task fallback for an automated driving system whose ability to monitor the driving environment has been compromised
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 safety challenges faced by Automated Driving Systems (ADS) when their perception modules fail or are compromised by hazardous weather, particularly in environments where stopping the vehicle is dangerous, such as tunnels or highways without hard shoulders. The authors argue that while lower-level automation relies on human intervention, automation-induced drowsiness and hypo-vigilance often prevent drivers from responding effectively to takeover requests. Consequently, the study proposes a Dynamic Driving Task (DDT) fallback strategy for Level 4 and Level 5 ADS, enabling the system to maintain safe operation without relying on human input or vehicle-to-infrastructure communication. The research was conducted using the SCANeR Studio driving simulation software, modeling a 2 km dual carriageway (Dampierre Road) characterized by limited visibility due to trees and potential adverse weather. The experimental design involved four vehicles, including an ADS-equipped vehicle, subjected to simulated perception failures in specific zones. The authors first analyzed the risks of having no fallback strategy, demonstrating that an ADS ignoring missing obstacles would collide with leading vehicles, while sudden deceleration upon recovery could cause multi-vehicle crashes. To mitigate these risks, the proposed strategy assumes the localization module remains functional, allowing the vehicle to stay in its lane. The core of the strategy involves a speed profile constrained by a Time to Collision (TTC) criterion of 4 seconds, derived from an embedded visibility map that calculates the maximum distance at which the ADS vehicle can be seen by following traffic. Speed limits were set between 20 km/h and 35 km/h to balance collision mitigation with traffic flow. The results indicate that the proposed fallback strategy successfully maintains a minimal risk condition. By replacing missing obstacles in the world model with "ghost objects" based on the last known state for one second, the system initiates preventive braking to smoothly transition to the TTC-compliant speed profile. This approach ensures the ADS vehicle does not exceed 30 km/h, thereby mitigating rear-end collisions with vehicles ahead. Furthermore, the 4-second TTC criterion guarantees that following vehicles have sufficient time to react, avoiding severe collisions from behind. The simulation confirmed that the ADS could traverse the hazardous road section in under six minutes without stopping, adhering to safety constraints despite the total loss of environmental perception. The significance of this work lies in providing a robust, sensor-independent fallback mechanism for high-level automation systems. It demonstrates that ADS can achieve a minimal risk condition through careful speed management and localization, even when perception is entirely compromised. This approach is critical for the deployment of Level 4 and 5 autonomous vehicles in complex road environments where emergency stops are prohibited, offering a solution that does not rely on the unpredictable behavior of human drivers or vulnerable communication networks.
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 | 2026-06-20 |
| archive | success | semantic_scholar | — | — | 6 | 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 | 2026-06-20 |
| 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