Capturing Driver Status Using Naturalistic Driving Study Data: Implications for Assessing Health and Safety

Abstract

The design of advanced driver-assistance systems (ADASs) and driving monitoring systems can be improved with a better understanding of on-road driving behavior. Information on traffic environment, health, engagement in secondary tasks, and corresponding eye glance locations can provide important insights into driver’s in-vehicle movements and status. This study examines the in-vehicle head movement of the driver and the factors affecting it using data from a naturalistic driving study. Linear mixed models were used to examine head position at a given point in time, and hierarchical clustering with dynamic time warping (DTW) was used to explore the trajectories of the head position. The findings show that some health conditions (e.g., limited flexibility, cancer) were correlated to restrictive head position during driving, and vehicle acceleration and traffic environment factors contributed to the variation in head positions.