Naturalistic driving data and driver headlighting needs

Michael J. Flannagan

The University of Michigan, Ann Arbor, Michigan USA

Contact email address: mjf@umich.edu

Abstract

Naturalistic driving data are now available to allow detailed measurement of how people drive at night. In naturalistic driving studies, data are collected by using instrumented vehicles to observe the driving behavior of normal drivers going about their usual daily activities. This provides good information about the circumstances of real traffic. Data are often collected over an extended period, perhaps several weeks, and the extended period helps to capture a range of circumstances. With instrumentation, vehicle movement can be characterized in detail in terms of speed, curvature, relationship to roadways and to other vehicles. This detailed information can then be used to understand the needs that drivers have to see various locations and distances on the roadway ahead of them. Analyses of recent naturalistic driving data are presented, and the relationship of these data to crash data and headlamp illumination are discussed. The results can guide the development of innovative adaptive forward lighting systems, such as Adaptive Driving Beam (ADB), but would also be useful for optimizing traditional static headlamps.