Cooperation Between Vehicle and Driver: Predicting the Driver’s Takeover Capability in Cooperative Automated Driving Based on Orientation Patterns
DOI: 10.1007/978-3-031-60494-2_17
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 ensuring safe transitions between automated and manual driving in cooperative automated driving systems. The core problem is that current Driver State Monitoring Systems (DSMS) cannot reliably predict a driver’s future takeover capability based solely on observation during automated driving. To enable smooth cooperation, the vehicle and driver must share a mutual understanding of their respective capabilities, conceptualized here as "confidence horizons." Since predictive criteria for takeover readiness are currently unavailable, the authors propose an alternative method called the Diagnostic Takeover Request (Diagnostic TOR). This approach predicts takeover capability by analyzing the driver’s initial orientation reaction—specifically gaze patterns—immediately after a takeover request is issued, rather than inferring state from behavior during the automated phase. The study validates this concept through two driving simulator experiments conducted at RWTH Aachen University. In both studies, participants drove at SAE Level 2 automation speeds (130 km/h) on a highway scenario. Half of the participants engaged in a non-driving-related task (playing Tetris) to induce variance in attention levels. A critical situation involving a broken-down vehicle triggered a takeover request. The researchers recorded gaze sequences using eye-tracking technology, defining Areas of Interest (AOIs) such as the road ahead, instrument cluster, and rearview mirrors. The first study involved 50 subjects using a head-mounted tracker, while the replication study involved 38 subjects using a remote camera system to ensure generalizability. Takeovers were classified as successful (safe braking or lane change) or unsuccessful (collision or unsafe maneuver). The results demonstrated that distinct gaze patterns precede successful and unsuccessful takeovers. Analysis of the combined data from both studies revealed that specific orientation sequences significantly correlated with takeover outcomes. For instance, gaze patterns initiating at the TICS display (center console) were strongly associated with unsuccessful takeovers, likely because drivers focused on the warning source rather than the road environment. Conversely, patterns involving immediate orientation toward the road or mirrors were more common in successful takeovers. The data showed that even with small sample sizes per specific gaze sequence, the likelihood of an unsuccessful takeover could be predicted with high probability for certain patterns (e.g., 100% likelihood for specific sequences like "T" or "IC"). This confirms that orientation reactions differ significantly before safe and unsafe takeovers. The significance of these findings lies in the potential to reduce reaction time during critical transitions. By detecting risky takeovers during the initial orientation phase, the system can initiate safeguarding measures, such as a Minimum Risk Maneuver, before the driver fully intervenes. This approach offers a practical, immediate application for DSMS using existing camera technology, bypassing the need for complex physiological sensors. The authors conclude that while gaze direction is only one component of driver capability, the Diagnostic TOR provides a viable method for early detection of takeover risks. Future work aims to integrate additional variables, such as body posture and grip strength, into the confidence horizon model to enhance the precision of cooperative control between vehicle and driver.
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 | Crossref | — | — | 1 | 2026-06-06 |
| archive | success | canonical_url | — | — | 1 | 2026-06-09 |
| extract | success | cached | — | — | 2 | 2026-06-09 |
| clean | success | clean | — | — | 1 | 2026-06-07 |
| chunk | success | chunk | — | — | 1 | 2026-06-07 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-07 |
| promote | success | — | — | — | 1 | 2026-06-06 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-09 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | success | — | — | — | 1 | 2026-06-09 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-09; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
- automation
- takeover transitions
- situational awareness
- automation surprise
- mode awareness
- odd communication
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: measurement protocol
- Theoretical Contribution: conceptual framework, computational model