An Analysis of Driver-Initiated Takeovers during Assisted Driving and their Effect on Driver Satisfaction
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 study investigates the reasons behind driver-initiated takeovers during the use of Advanced Driver Assistance Systems (ADAS) and evaluates their impact on driver satisfaction. While system-initiated takeovers are well-documented, driver-initiated interventions in naturalistic driving remain poorly understood. The authors aim to categorize these interventions and determine if they can serve as feedback for ADAS optimization and individualization, specifically focusing on a SAE Level 1 Predictive Longitudinal Driving Function (PLDF). The research employed a test group study involving 17 participants who used the PLDF for their daily commutes over one week in southwestern Germany. Participants drove equipped Porsche vehicles, utilizing the system for longitudinal control while manually steering. Crucially, drivers were required to annotate every intervention using voice recordings, specifying the driver input, situation, reason, and desired behavior. This approach provided ground-truth data on intervention motives, overcoming limitations of previous studies that relied solely on vehicle data. The resulting dataset comprised 165 drives, totaling 92.8 hours and 4,334 km, with 3,335 annotated interventions. Post-study questionnaires assessed driver satisfaction using a Likert scale. The analysis revealed three primary categories for driver-initiated takeovers. The most frequent category, accounting for 53.7% of interventions, stemmed from personal preferences deviating from the system’s default behavior, such as adjusting speed on straight roads or altering acceleration timing. A second category involved corrections for incorrect sensing or map data (12.3%), while the third comprised necessary interventions due to system limitations outside its Operational Design Domain (ODD), such as traffic interactions (27.7%). Correlation analysis demonstrated a significant negative relationship between driver satisfaction and both the number and frequency of interventions, particularly those occurring within the ODD. Interventions outside the ODD did not significantly affect satisfaction, suggesting drivers accept these as inherent system limitations. The findings indicate that a majority of interventions are preventable through better alignment with driver preferences and improved data accuracy. Since reducing interventions within the ODD significantly increases satisfaction, the study concludes that driver takeover behavior serves as valuable feedback for ADAS optimization. Furthermore, substantial variations in intervention patterns among individuals highlight the necessity for ADAS personalization. The authors suggest that future systems should utilize this intervention data to iteratively adapt to individual driving styles, thereby enhancing user acceptance and satisfaction.
Key finding
In a naturalistic SAE Level 1 ADAS commuter study, ~54% of driver-initiated takeovers occurred inside the system's Operational Design Domain to align ADAS behavior with personal preference, and only within-ODD takeover counts and frequencies (not out-of-ODD ones) were significantly correlated with reduced driver satisfaction, motivating individualized ADAS adaptation driven by intervention feedback.
Methodology
naturalistic
Sample size: 17
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 discover_arxiv on 2026-05-03 (2 acquisition events logged).
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | arxiv | — | — | 3 | 2026-05-03 |
| archive | success | — | — | — | 1 | 2026-05-03 |
| 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 | 2026-05-03 |
| promote | success | — | — | — | 1 | 2026-05-03 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 2 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 15 | 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).
- Empirical Findings: behavioral performance data
- Theoretical Contribution: conceptual framework, computational model