Physiological Indicators of Driver Workload During Car-Following Scenarios and Takeovers in Highly Automated Driving

Radhakrishnan, Vishnu; Merat, Natasha; Louw, Tyron; Goncalves, Rafael; Lyu, Wei; Torrao, Guilhermina; Guillen, Pablo Puente; Lenné, Michael G. · 2021 · Crossref

DOI: 10.31234/osf.io/bwp6f

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 car-following scenarios and takeover events in highly automated driving (HAD) affect driver workload, using physiological indicators to objectively measure cognitive and attentional demands. The research addresses the "out-of-the-loop" phenomenon, where drivers disengage from monitoring the environment during automation, potentially leading to performance decrements when resuming control. Specifically, the paper examines how Time Headway (THW) and engagement in non-driving related tasks (NDRTs) influence workload during automated car-following (ACF), manual car-following (MCF), and takeover phases. The experiment utilized a driving simulator with 32 participants divided into two groups: SAE Level 2 (L2), where drivers monitored the drive, and SAE Level 3 (L3), where drivers engaged in an NDRT (an "Arrows Task") during automation. Participants completed two ~18-minute experimental drives featuring Short (0.5 s) and Long (1.5 s) THW conditions. Data collection included Electrocardiogram (ECG) and Electrodermal Activity (EDA) signals, processed to derive metrics such as heart rate variability (RMSSD), heart rate, respiration rate, and skin conductance responses (SCRs), alongside self-reported workload ratings. Results indicated that driver workload during ACF was significantly higher in the L3 group (engaged in NDRT) compared to the L2 group (monitoring only), as evidenced by increased physiological activation across all ECG and EDA metrics. In contrast, THW conditions (Short vs. Long) did not significantly affect workload during the monitoring phase for L2 drivers. However, during takeover scenarios, the presence of a lead vehicle maintaining a shorter THW significantly increased driver workload, particularly affecting the L3 group. The study found that ECG and EDA signals were sensitive to these workload variations, with EDA metrics (specifically SCR frequency) being particularly responsive to the short-duration stress of takeover events. The findings suggest that while THW alone does not drastically alter workload during passive monitoring, the combination of NDRT engagement and short THWs during takeovers poses a significant cognitive load, potentially compromising safety. The study concludes that combining ECG and EDA signals offers a robust method for real-time workload assessment, which could inform the design of driver monitoring systems that predict driver readiness and mitigate performance drops during transitions from automated to manual control.

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.

• hands on hands off engagement
• workload measurement
• takeover transitions
• automation
• stress driving
• mental demand

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: physiological data, behavioral performance data
• Theoretical Contribution: theory or model