Citation

Levison W, Kantowitz BH. Transp. Hum. Fact. 2000; 2(3): 265-278.

Affiliation

William H. Levison Associates, Lexington, MA; University of Michigan Transportation Research Institute, USA

Copyright

(Copyright © 2000, Informa - Taylor and Francis Group)

DOI

10.1207/STHF0203_6

PMID

unavailable

Abstract

Field studies of driver behavior that are conducted to determine the operational effects of highway geometry typically involve measuring the speeds of many drivers at a few locations using roadside measuring equipment. In this article, we describe a methodology for obtaining speed estimates at many locations for a single driver using in-vehicle measuring equipment. Specifically, data from the Global Positioning System are used to define geographic coordinates of the highway and, in combination with on-board speed measurements, to provide near-continuous information of the location and speed of an instrumented test vehicle. To understand fully the impact of a particular segment of roadway geometry (e.g., a specific horizontal curve) on driver behavior, it is necessary to estimate the impact of the geometry before and after the test site. We propose an algorithm to predict whether or not a preceding curve will influence the maximum speed on the approach to a test site. This prediction is based on assumed driver characteristics-and to some extent vehicle characteristics-as well as on the road geometry.