Train driving simulator studies: can novice drivers deliver the goods?

Large, David R. and Golightly, David and Taylor, Emma (2017) Train driving simulator studies: can novice drivers deliver the goods? Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit . ISSN 2041-3017 (In Press)

[img] PDF - Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (583kB)

Abstract

Early research suggests that, in a simulated train-driving environment, unskilled, novice drivers may exhibit comparable behaviour and performance to experienced, professional train drivers after receiving only minimal, task-specific training. However, this conclusion is based on exiguous performance indicators, such as speed limit exceedances, SPAD violations etc., and considers only limited data. This paper presents further, detailed analysis of driving performance data obtained from 20 drivers (13 novices and 7 experienced train drivers), who took part in a previous simulator-based research study, utilising more sensitive and perspicuous measures, namely acceleration noise and control actuation. Results indicate that, although both cohorts exhibited similar performance using the original metrics, and would thus support the same conclusions, the manner in which this performance was effected is fundamentally different between groups. Trained novice drivers (mainly comprising students and staff at the University of Nottingham) adopted far more erratic speed control profiles, characterised by longer control actions and frequent switching between power and brake actuation. In contrast, experienced drivers delivered smoother acceleration/braking profiles with more subtle (and shorter) control actions and less variance in speed. We conclude that although utilising trained non-drivers may offer an appealing solution in the absence of professional train drivers during simulator-based research, and their input remains of value, researchers should remain mindful when interpreting results and drawing conclusions from a contingent comprising non-drivers. The work also demonstrates the value of dependent variables such as acceleration noise, and quantitative measures of control actuation, which may offer an insightful portfolio of measures in train driving research studies.

Item Type: Article
Schools/Departments: University of Nottingham, UK > Faculty of Engineering
Depositing User: Eprints, Support
Date Deposited: 07 Apr 2017 09:46
Last Modified: 08 Apr 2017 06:29
URI: http://eprints.nottingham.ac.uk/id/eprint/41819

Actions (Archive Staff Only)

Edit View Edit View