No single camera position captures fatigue. The driver, the vehicle and the environment are each partial sensors. Rotating mount positions across days surfaces signals that remain hidden in any one fixed view.
iPhone 14+ with 60fps video, accelerometer and gyroscope streams. Mounted on suction arm or dash cradle, rotated per schedule. Supplementary audio capture throughout.
Driver-facing or road-facing video at 60fps, 1080p minimum. Rotated per schedule.
Accelerometer and gyroscope from mounted iPhone — picks up steering micro-corrections, braking smoothness, road texture.
Continuous audio capture — yawns, sighs, vocal tonality changes, ambient cabin noise and engine load.
Timestamped GPS + road type log for traffic density, route complexity, glare and monotony context.
Participants complete a short structured survey before every driver journey, capturing subjective baseline state: sleep duration and quality the previous night, Karolinska Sleepiness Scale rating, caffeine and meal intake in the past two hours, mood, and any medication or illness flags.
Participants complete a matching survey immediately after every driver journey: end-state KSS, perceived fatigue and effort, critical incidents or near-misses, route conditions (traffic, weather, glare) and any unplanned stops. Bookending every drive gives us delta between expected and experienced state.
Each participant wears an issued Fitbit 24/7 for the full study period, including sleep, and is responsible for charging it daily (shower or fixed routine slot). Captures continuous heart rate, HRV, sleep stages and consistency, activity and sedentary periods, SpO₂ and skin temperature — the physiological spine beneath the observational data.
Drivers are briefed on: wearing the Fitbit at all times and syncing daily; completing both surveys for every journey (no exceptions); charging the device on a fixed daily routine; flagging any missed wear period longer than four hours. Non-compliance is logged, not punished — but compliance is the dataset.