Quantifying drivers’ comfort-zone and dread-zone boundaries in left turn across path/opposite direction (LTAP/OD) scenarios

AbstractThe aim of this study is to quantify drivers’ comfort- and dread-zone boundaries in left-turn-across-path/opposite-direction (LTAP/OD) scenarios. These scenarios account for a large fraction of traffic fatalities world-wide. The comfort zone is a dynamic spatiotemporal envelope surrounding the vehicle, within which drivers feel comfortable and safe. The dread zone, a novel concept, describes a zone with a smaller safety margin that drivers will not voluntarily enter, but can push themselves into when conditions provide additional motivation (e.g., when hurried). Quantifying comfort- and dread-zone boundaries in the context of turning left before or after an oncoming vehicle has the potential to inform and improve both the design and driver acceptance of advanced driver assistance systems (ADAS) and autonomous vehicles.Using a within-subject design, a test-track experiment was conducted with drivers turning an instrumented vehicle left across the path of an oncoming vehicle. The oncoming vehicle was a self-propelled full-sized computer-controlled balloon vehicle going straight at a constant speed (50km/h). The driver assumed full control of the instrumented vehicle approximately 20m before the intersection and had to make the decision to turn left before or after the oncoming balloon vehicle. There were two experimental conditions, comfortable driving and hurried driving. Measures for each turn included post-encroachment time (PET), lateral acceleration, and self-reports of comfort and risk. Drivers consistently accepted shorter time gaps and higher lateral accelerations when hurried. We interpret these findings to suggest that drivers invoke two dynamic, contextually-defined safety margins. The first is the comfort-zone boundary, a limit which drivers do not voluntarily cross without extra motives. The second is the dread-zone boundary, a more distant limit which drivers do not voluntarily cross even with extra motives. Grouping the responses (high/low) to the driver behavior questionnaire (DBQ) improved the ability to predict the dread-zone boundary PET given the comfort-zone boundary PET