Effects of non-driving-related tasks on driving performance after takeover request in automated driving (Level 3): a simulator study

Horáková, Miroslava; Vondráčková, Lucie; Krejčí, Libor; Trepáčová, Martina · 2025 · Crossref

DOI: 10.1186/s12544-025-00745-6

archive: archived pipeline: cataloged

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

Summary

This study investigates how different categories of non-driving-related tasks (NDRTs) affect driver performance during the takeover process in Level 3 (conditional) automated driving. The research addresses the critical safety concern that drivers, while disengaged from manual control, often engage in secondary activities that may impair their ability to promptly and safely assume vehicle control when a takeover request (TOR) is issued. The primary objective was to evaluate the specific impact of six distinct NDRTs on driving metrics following a TOR, using a high-fidelity truck and bus simulator to simulate realistic freight transport scenarios where such automation is expected to be prevalent. The experimental design involved 31 professional drivers (final analysis sample: 28 males, 3 females; mean age 40) who participated in three sessions: a training session, a baseline driving session without NDRTs, and an experimental session where six NDRTs were performed for 10 minutes each before a TOR was triggered. The six tasks were: working on a laptop, making a phone call, watching a video, taking an audio course, relaxation with closed eyes, and a manual assembly task. These tasks varied in sensory modality (visual vs. auditory) and cognitive load. Driving performance was measured using objective metrics including takeover time, average speed, lateral deviation from the lane center, and situational awareness assessed via the SAGAT method. Statistical analysis used the Wilcoxon Signed-Rank test to compare performance against the baseline. Results indicated that all NDRTs impaired driving performance to some degree, but the magnitude of impairment varied significantly by task type. Working on a laptop, watching a video, and making a phone call were identified as the riskiest activities. Statistically significant increases in takeover time were observed for relaxation, working on a laptop, manual tasks, and watching a video, with the manual task yielding the longest average takeover time (4.81 seconds) compared to the baseline (approx. 2.6–3.0 seconds). Significant decreases in average speed were found for relaxation, laptop work, audio courses, manual tasks, and video watching. However, no significant differences were found in lateral distance to the lane center for any task. Situational awareness was lowest during phone calls (58.08% accuracy) and highest during video watching (98.38%), suggesting that visual tasks may maintain better environmental monitoring than auditory ones, despite their potential for distraction. The findings suggest that the sensory modality and cognitive load of an NDRT are key determinants of takeover performance. Visual tasks, particularly those with high mental load like laptop work, significantly delay takeover and reduce speed, whereas auditory tasks like phone calls, while distracting, did not significantly prolong takeover time in this context. The study concludes that not all NDRTs are equally safe; drivers should be cautioned against engaging in high-load visual tasks during automated driving phases, as these significantly compromise the ability to react to takeover requests. These results provide evidence for designing driver-vehicle interfaces that may restrict or warn against specific high-risk activities during Level 3 automation.

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 qwen3.8-27b-gittensor on 2026-08-23; verification: pending re-verification.

Topics

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

• takeover transitions
• manual
• automation
• temporal
• visual manual
• automation complacency bias

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