Levels of train automation: classification and determining their impact on driver mental workload and vigilance task performance

Download files
Access & Terms of Use
open access
Copyright: Spring, Peter Anthony
Altmetric
Abstract
Three experiments were conducted to investigate the impact of: (1) four Levels Of Automation (LOAs) on driver vigilance, (2) two vigilance device types on driver mental workload, at two LOAs, and (3) two vigilance device types on driver vigilance, at two LOAs. In experiment 1, student participants drove a train simulator with one of four levels of automation: manual (nil), in-cab signalling (low), ATP (medium), and an autopilot (high). Results revealed that driver vigilance was negatively impacted by train automation, but only at a high level (i.e., an autopilot). This finding was interpreted as relating to driver mental workload. That is, a high LOA may have resulted in drivers being mentally underloaded. This interpretation required empirical validation in the following two experiments. In experiment 2, the task demands of participants in two high LOA groups (suspected of being mentally underloaded) were increased to one of two levels by means of a secondary task. Participants in two low LOA groups had the same treatments, and served as comparison groups. The secondary task was either a sensory vigilance device (low additional task demands) or a cognitive vigilance device (high additional task demands). Results revealed that autopilot supervisor (high LOA) participants were indeed underloaded, even with the sensory vigilance device secondary task. However, autopilot supervisor participants with the cognitive vigilance device secondary task were not underloaded. On the other hand, the in-cab signalling (low LOA) group of participants appeared to be overloaded when concurrently operating the cognitive vigilance device, but not when operating the sensory vigilance device. In experiment 3, the negative impact of a high LOA on driver vigilance was predicted to be counteracted by the additional (high) task demands of the cognitive vigilance device, but not the additional (low) task demands of the sensory vigilance device. This prediction was supported by the results in terms of the cognitive vigilance device. However, the sensory vigilance device also counteracted the negative impact of a high LOA. An explanation for this was in terms of a startling effect of the sensory vigilance device, rather than it being due to increased task demands. Future studies were suggested that could determine if mental overload still occurs when a verbal/auditory, and/or a task-linked, version of the cognitive vigilance device is used.
Persistent link to this record
Link to Publisher Version
Link to Open Access Version
Additional Link
Author(s)
Spring, Peter Anthony
Supervisor(s)
McIntosh, Andrew
Caponecchia, Carlo
Creator(s)
Editor(s)
Translator(s)
Curator(s)
Designer(s)
Arranger(s)
Composer(s)
Recordist(s)
Conference Proceedings Editor(s)
Other Contributor(s)
Corporate/Industry Contributor(s)
Publication Year
2011
Resource Type
Thesis
Degree Type
PhD Doctorate
UNSW Faculty
Files
download whole.pdf 3.86 MB Adobe Portable Document Format
Related dataset(s)