An Examination of Overtaking Judgments Based on Limitations in the Human Perceptual System: Implications for the Design of Driver-Assistance Systems - Digital Signal Processing for In-Vehicle Systems and Safety - page 225

17.1 Introduction

Traffic accidents account for several deaths and injuries each year. In Europe,
42,500 people are killed and 3,500,000 are injured every year due to traffic
accidents [ 1 ]. Similarly, in the United States, over 42,000 people are killed in a
year due to motor vehicle crashes [ 2 ]. Therefore, it is important to understand the
underlying reasons for such collisions. It has been suggested that approximately
90% of all traffic accidents can be attributed to human error [ 3 ], and studies have
investigated driver performance during car following [ 4 ], left and right turn
maneuvers, [ 5 ] and at traffic intersections [ 6 ]. Also, research on computer vision
and intelligent transportation systems have led to the development of several driver-
assistance systems (DAS) that aid drivers, for example, in car following, inattention
detection, pedestrian spotting, and lane keeping [ 7 - 9 ].
However, fewer studies have focused on drivers' judgments during overtaking
maneuvers [ 10 , 11 ]. This is quite surprising because overtaking maneuvers lead to
many fatal accidents each year [ 12 ]. For example, between 1995 and 2000, about 26
traffic participants died each year in Netherlands due to overtaking failures [ 12 ].
In addition, it was reported that overtaking maneuvers led to a considerable propor-
tion of injury-causing accidents in Nottinghamshire, England [ 13 ]. Furthermore, in
the United States, there were 138,000 accidents due to overtaking in the year 2000,
and such accidents accounted for 2.1% of all fatal crashes and 1.1% of injury
crashes [ 14 ]. In short, global accident data suggests that it is critical to identify the
underlying causes for accidents during overtaking maneuvers. Since driving is
primarily a visual task [ 15 ], it will be beneficial to identify limitations in the
human perceptual system that lead to erroneous judgments during such maneuvers.
Consequently, it is essential to explore ergonomically appropriate solutions to
overcome such limitations by developing DAS to help drivers during overtaking
maneuvers.
In this article, we address five specific topics. First, we identify the sources of
visual information that drivers rely on during an overtaking maneuver and the
perceptual judgments that they typically make during such maneuvers. Second, to
better understand the complexity of an overtaking maneuver, we compare and
contrast an overtaking task with car following, especially because judgments during
car following have been widely studied [ 16 - 18 ]. We outline the critical differences
in the available visual information and associated judgments during the two types of
tasks. Third, since forward-collision-avoidance-warning-systems (FCAWS) are
available to aid drivers during car following, we investigate the possibility of
using typical FCAWS to aid overtaking maneuvers. Based on the known limitations
in the human perceptual system and the functional capabilities of the currently
available FCAWS, we report certain disadvantages in using such FCAWS to
support overtaking maneuvers. Fourth, we describe seven functional requirements
that are important to be considered in the design of an ergonomically efficient DAS
to support overtaking maneuvers. Finally, we propose a model for the design of
DAS that emphasizes on overcoming drivers' perceptual limitations by enhancing
the effectiveness of the available visual information.