Perceptual Errors in Judging the Approach of Motor Vehicles. / Gould, Mark.

2013. 214 p.

Research output: ThesisDoctoral Thesis

Unpublished

Abstract

Motorcycles are vastly overrepresented in road accident statistics across the world, with 61% of all car and 75% of all motorcycle accidents occurring at road junctions (Department for Transport, 2010a). When drivers attempt to negotiate their way safely out of a road side junction and into the flow of traffic, the most dependable cues for judging whether an approaching vehicle poses an immediate threat are its optical size and its rate of expansion on the retina (Lee, 1976). While this information may appear to be the most reliable, research has demonstrated that individuals gauge the time-to-passage (TTP) of smaller vehicles less accurately than larger vehicles (e.g. Caird & Hancock, 1994). This thesis investigated the perceptual mechanisms that underlie driver’s abilities to make judgements about the immediacy of the threat posed by approaching vehicles at roadside junctions. This is investigated in three areas; judgements of relative speed, detection of vehicle approach and the effect of conspicuity aids.
The first experimental chapter explored decrements in judgements of motorcycle approach speed when only the white headlight is available as a cue on a black background, and how accuracy is improved by adding two flanking lights to the solo headlight in order to create a triangular headlight arrangement. The chapter also investigated the optimal configuration of the tri-headlight arrangement on the accuracy of approach speed judgements. In the following chapter, participants gauged motorcycle and car speeds within a virtual city scene as the ambient light level was manipulated. The study demonstrated that the enhancement in
4
performance through the use of the tri-headlight occurred once the contour of the motorcycle and motorcyclist could no longer be differentiated from the background, whereas decreasing ambient light level did not affect car speed judgements.
In Chapter 5, the thesis progressed to investigate the ability of individuals to detect whether a motorcycle or car was approaching their viewpoint within a virtual city scene. Individuals displayed significantly poorer thresholds for detecting an approaching motorcycle compared with an approaching car. Additional foveal motion caused a significant decrement in detection thresholds for cars but not motorcycles, although this is likely to be due to a ceiling effect for motorcycle detection. In Chapter 6, the role of additional motion was investigated further as thresholds for the detection of vehicle approach were assessed in the presence of simulated self-motion, which lead to significant impairment in detection thresholds for cars and motorcycles. In Chapter 7, the effect of a high visibility vest on detection thresholds was investigated, but no significant effects were found.
Overall, the thesis demonstrates the limitations of the human perceptual system in judging the relative speed of a motorcycle compared with a car stimulus, a problem which is exacerbated under low levels of luminance. However, the simple engineering solution of additional headlights is shown to vastly improve these speed judgement impairments under low levels of luminance. The thesis provides
5
evidence that individuals are less sensitive to the detection of motorcycle approach compared with a car stimulus. The effect of additional scenic motion is shown to negatively affect car detection sensitivity, while simulated self-motion is shown to impair detection thresholds for motor vehicles. Implications for road design are discussed. Lastly, the thesis demonstrates that high visibility garments do not significantly improve detection capabilities for motorcycles.
Original languageEnglish
QualificationPh.D.
Awarding Institution
Supervisors/Advisors
  • Wann, John, Supervisor
  • Helman, Shaun, Supervisor, External person
Award date1 Jun 2013
Publication statusUnpublished - 2013
This open access research output is licenced under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.

ID: 17157852