Getting back in the loop: Does autonomous driving duration affect driver's takeover performance?

Portron, Arthur; Perrotte, Gaëtan; Ollier, Guillaume; Bougard, Clément; Bourdin, Christophe; Vercher, Jean-Louis · 2024 · Crossref

DOI: 10.1016/j.heliyon.2024.e24112

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates how the duration of Level 3 autonomous driving affects a driver’s ability to safely regain control of the vehicle upon a takeover request (TOR). The research addresses the "out-of-the-loop" phenomenon, where drivers disengage cognitively and physically during automated phases, potentially impairing their situation awareness and reaction capabilities when manual intervention is required. While previous studies often examined short automation periods, this work specifically tests the hypothesis that takeover performance degrades linearly as autonomous driving duration increases, using four distinct time intervals: 5, 15, 45, and 60 minutes. The experiment involved 52 participants divided into four independent groups, each assigned one of the specified autonomous driving durations. Using a static driving simulator, participants engaged in a monotonous highway driving scenario. During the autonomous phase, they performed a non-driving-related task (watching a movie) to simulate realistic disengagement. Each participant completed two identical experimental sequences separated by a 10-minute break to assess adaptation effects. In each sequence, a TOR was issued when the lead vehicle suddenly braked, providing an 8.3-second time budget to avoid a collision. Performance was measured using oculomotor data (time to first eye on road), reaction times, and kinematic metrics such as minimum time-to-collision, maximum deceleration, and lateral deviation during the avoidance maneuver. Drowsiness levels were also monitored via video analysis. The results rejected the initial hypothesis of a linear relationship between automation duration and performance degradation. Instead, a nonlinear pattern emerged, with the 15-minute group demonstrating the safest overall performance. Performance varied significantly across groups, with mixed outcomes in longer duration conditions. Crucially, the study found that repetition induced a major improvement in performance during the second drive, indicating rapid adaptation to the takeover scenario. The non-driving-related task was identified as a critical factor influencing driver engagement and readiness. These findings suggest that the impact of autonomous driving duration on takeover performance is complex and not strictly proportional to time spent out of the loop. The superior performance in the 15-minute group and the significant adaptation effect imply that driver readiness is influenced by factors beyond mere duration, such as task dynamics and familiarity with the system. The study supports reservations about treating automated driving similarly to passive transport like train travel, emphasizing the need for strategies to maintain driver situation awareness and facilitate rapid re-engagement.

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.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• takeover transitions
• automation
• time on task
• temporal
• manual
• automation surprise

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
• Methodological Resource: measurement protocol
• Theoretical Contribution: conceptual framework