A validation study comparing performance in a low-fidelity train-driving simulator with actual train driving performance

Abstract

BackgroundAlthough common in other industries, such as the automotive sector, no train-driving validation study has been found in the existing literature. The present paper intends to fill that gap by comparing the results of train-driving performance in a physically low-fidelity but highly functional simulator with real train-driving performance. MethodThirty-four train driver students in the final part of their basic education were assessed in a 45-minute simulator test using the number of driving errors as the performance indicator. The results were compared with the performance at 11 weeks of internship as measured by supervisors grading according to a standard procedure. One of the classes (17 to-be drivers) was affected by restrictions related to COVID-19, which led to a shortened internship and distance learning during parts of the internship. The study also intended to measure the effect of the restrictions and the types of errors the drivers made by comparing the two classes. ResultsA significant correlation was found between the number of driving errors and internship grades, r = −0.45, p <.05. The results also revealed that COVID-19 restrictions negatively affected performance, as the students from Class B made significantly more driving errors and obtained a lower internship grade than those from Class A. ConclusionsThis paper shows that this type of low-fidelity simulator is well suited for measuring real train-driving performance. A measurement method that can predict long-term driving should have implications for both research and practical usability. Researchers can use this for studying the effects of, for example, different training methods, while train operation companies can use the method to test their drivers' skills and intervene before an actual accident occurs.