Drivers’ engagement in NDRTs during automated driving linked to travelling speed and surrounding traffic

Liu, Xian; Madigan, Ruth; Sadraei, Ehsan; Lee, Yee Mun; Merat, Natasha · 2023 · Crossref

DOI: 10.31234/osf.io/mhwrp

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 driving environment factors, specifically travelling speed and surrounding traffic conditions, influence drivers’ engagement in Non-Driving Related Tasks (NDRTs) during SAE Level 3 automated driving. While previous research indicates that drivers become more inattentive and engage more in NDRTs during automation compared to manual driving, it remains unclear how real-world traffic dynamics affect this behavior in Level 3 vehicles, where drivers are legally permitted to disengage from the driving task. Understanding these patterns is critical for assessing driver readiness to resume control and for designing effective human-machine interfaces. The researchers conducted an on-road study using a Level 3 automated test vehicle on a 95 km European motorway. Seventy-nine non-professional drivers participated, but the final analysis focused on 46 video clips from 32 drivers who actively engaged in NDRTs for at least one minute during automation. Due to the absence of usable external cameras, the study used vehicle CAN bus data to proxy traffic conditions: mean speed represented traffic volume (lower speed indicating higher density), and the standard deviation (SD) of speed represented traffic fluctuation. Video recordings were coded to measure NDRT engagement via the number and mean duration of glances away from the NDRT toward the road or dashboard per minute. Generalised Linear Mixed Models (GLMMs) were employed to analyze the relationships between these environmental proxies, demographic variables (age, gender, prior AV experience), and glance behaviors. The results demonstrated that traffic dynamics significantly impacted visual attention. A higher SD of speed, indicating frequent acceleration and deceleration due to traffic fluctuations, led to a significant increase in both the number and mean duration of glances away from NDRTs. This suggests drivers monitored the environment more closely when the vehicle’s motion was unstable. Additionally, lower mean speeds were associated with longer glance durations away from NDRTs, implying that drivers in congested, slow-moving traffic paid more sustained attention to the roadway than those in free-flowing high-speed traffic. Demographic factors also played a role: older drivers glanced away from NDRTs more frequently, while male drivers exhibited both more frequent and longer glances away from NDRTs compared to female drivers. Prior AV experience did not significantly influence engagement levels. These findings indicate that drivers self-regulate their attention based on the complexity and stability of the driving environment, even during Level 3 automation. Contrary to some manual driving studies where low speed encourages distraction, this study found that low speed and high speed variability prompted increased monitoring of the driving environment. This highlights that drivers remain sensitive to traffic cues and potential hazards, allocating visual resources to maintain situational awareness. The results provide valuable insights for AV designers, suggesting that human-machine interfaces should account for these dynamic attention shifts to ensure drivers remain prepared to take over control when necessary.

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.

• temporal
• visual
• situational awareness
• automation
• external distraction
• hands on hands off engagement

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, observational prevalence
• Theoretical Contribution: conceptual framework