This dataset contains anonymized experimental data collected from a driving simulator study conducted at Seoul National University. The study examined how hazard type (behavioral vs. environmental) and takeover request (ToR) explanation type (action-only, precursor-only, hazard-only, precursor + hazard) jointly influence drivers’ cognitive and behavioral responses during conditionally automated driving (SAE Level 3).
Design: 4 (Explanation Type) × 2 (Hazard Type) within-subjects design
Hazard Types:
- Behavioral Prediction (BP)
- Environmental Prediction (EP)
(Based on Crundall’s hazard taxonomy distinguishing driver reasoning demands.)
Explanation Types:
- E1: Action-only
- E2: Precursor-only
- E3: Hazard-only
- E4: Precursor + Hazard
Participants: 12 licensed drivers (normal or corrected-to-normal vision/hearing)
Simulator: Forum 8 UC WinRoad ver. 10
Task: Participants engaged in non-driving-related smartphone tasks during automated driving and responded to system-issued takeover requests.
Takeover timing: ToR issued at 6,000 m with a 7 s takeover lead time and 10 s time-to-collision.
File format: CSV (UTF-8 encoded)
Each participant contributed 8 rows (4 explanation × 2 hazard conditions), totaling 96 rows.