2H1-6　The Effect of Driver Engagement in Autonomous Driving based on Flow Experience

Park, Jinsang; Jung, Eui S.; Im, Youngjae · 2019 · Crossref

DOI: 10.5100/jje.55.2h1-6

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 driver engagement and behavioral patterns in semi-autonomous driving environments, specifically focusing on how Non-Driving Related Tasks (NDRTs) influence driver state based on the four-channel model of flow theory. As autonomous technology shifts the driver’s role from operator to supervisor, understanding the interaction between driver skill, task challenge, and mental workload is critical for ensuring safe take-over performance. The research aims to determine how different combinations of NDRT difficulty and driver skill create distinct psychological states—Apathy, Relaxation, Anxiety, and Flow—and how these states affect reaction times and mental workload during take-over scenarios. The experiment involved 12 participants (aged 28–31) using a driving simulator configured for Level 3 automation. Participants performed two types of NDRTs: an Addition task (cognitive) and a Drawing task (motor), each with simple and complex difficulty levels. These tasks were designed to manipulate the challenge level, while motivation techniques were used to influence skill perception, thereby inducing the four flow conditions. During autonomous driving at 100 km/h, participants engaged in these tasks until a Take-Over Request (TOR) was issued. The study measured perceived demand and flow experience using the Flow Short Scale (FSS), mental workload using NASA-TLX, and reaction time defined as the interval between the TOR and the start of manual maneuvering. The results demonstrated that both NDRT type and flow condition significantly impacted driver metrics. Participants perceived the Addition task as having a higher demand level and inducing greater mental workload than the Drawing task. Regarding flow conditions, the Anxiety state resulted in the highest perceived demand and mental workload, while the Apathy state resulted in the lowest. The Flow condition yielded the highest flow experience scores but also the slowest reaction times (M = 3.665s). Conversely, the Apathy condition produced the fastest reaction times (M = 2.585s), though with the lowest engagement. The Relaxation and Anxiety conditions showed intermediate reaction times. Notably, the Flow condition was associated with an adequate, moderate level of mental workload, suggesting an optimal balance between engagement and cognitive load. The study concludes that the four-channel flow model effectively categorizes driver states in semi-autonomous driving, with distinct implications for safety and system design. While the Flow state represents an optimal user experience with balanced mental workload, it correlates with slower take-over reactions compared to low-engagement states like Apathy. This finding highlights a potential trade-off between driver satisfaction/engagement and immediate responsiveness. The results suggest that system designers must consider the type of NDRT and the induced psychological state when defining interaction guidelines, as cognitive tasks like addition impose higher loads and slower reactions than motor tasks like drawing. These insights contribute to developing intelligent systems that better manage the collaboration between human supervisors and automated vehicles.

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
• takeover transitions
• mental demand
• stress driving
• automation
• 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
• Theoretical Contribution: theory or model, conceptual framework