Driver Takeover Performance under Different Driving Contexts and Non-Driving Tasks in Level 3 Automated Driving

Liu, Meiqi; Zhu, Zhixian; Chen, Jie; Zhang, Yifan; Zhong, Jiaqian; Zhang, Yu · 2026 · Crossref

DOI: 10.54941/ahfe1007862

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 driving context and non-driving-related task (NDRT) demand jointly influence driver takeover performance in Level 3 (L3) automated driving. In L3 systems, drivers disengage from continuous control but must resume control upon a takeover request (TOR). The research addresses a gap in existing literature regarding the interactive effects of environmental complexity and NDRT cognitive load on takeover readiness. The authors hypothesized that urban environments, characterized by higher information density and conflict points, combined with high-demand NDRTs, would significantly impair takeover timing, attentional reallocation, and situational awareness. The researchers conducted a driving-simulator experiment with 18 licensed drivers using a 2 × 3 within-subject factorial design. The independent variables were driving context (highway vs. urban) and NDRT type (music listening, entertainment video viewing, and news summarization). Each participant completed six counterbalanced takeover trials. Takeover performance was evaluated using multimodal measures: takeover reaction time (TRT) annotated from video, eye-tracking metrics (time to first fixation and attention allocation), heart rate variability (HRV) indices (RMSSD, LF/HF), and subjective ratings of workload and situational awareness. Data were analyzed using repeated-measures ANOVA. Results indicated that both driving context and NDRT demand significantly affected takeover performance. Urban scenarios were associated with longer TRT, higher perceived workload, and more dispersed visual attention compared to highway scenarios. Among NDRTs, news summarization produced the poorest takeover outcomes, characterized by the longest TRT, greater off-road attention, slower gaze return after TOR, and reduced situational awareness. HRV data showed higher physiological stress (indicated by lower RMSSD and higher LF/HF) during high-demand tasks, particularly in urban conditions. Effect-size comparisons revealed that the driving context effect on TRT (Cohen’s dz = 0.683) was larger than pairwise differences among NDRTs. Additionally, more experienced drivers demonstrated shorter TRT in five of six conditions, suggesting experience buffers high cognitive load. The findings highlight that takeover performance is a context-sensitive process driven by attentional resource competition and task-switching costs. The combination of complex urban environments and visually/cognitively demanding NDRTs creates a high-cost transition for drivers. These results imply that L3 automated driving interfaces require context-aware takeover support that manages NDRT access based on environmental complexity and facilitates rapid attentional redirection. The study also suggests that drivers with limited experience may need stronger takeover assistance under high-load conditions. Limitations include reduced sample sizes after data exclusion and the reliance on manual video annotation for TRT, pointing to the need for larger samples and fully quantitative gaze analysis in future work.

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
• mental demand
• temporal
• automation
• manual
• situational awareness

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