Remote Control of Automation: Workload, Fatigue, and Performance in Unattended Railway Operation

Abstract

Human failure performance of staff working in increasingly automated work environments, particularly control rooms, has been reported to deteriorate along with related constructs such as workload and fatigue once degrees of automation increased. At the same time, mediation effects of domain‐specific contextual factors were stressed, questioning the generalizability of the proposed automation effects onto domain‐specific applications. This thesis examined whether a specific contextual factor—monotony—mediated the proposed detrimental effects of degrees of automation on workload, fatigue, and human failure performance in domain‐specific railway automation, described by the taxonomy of grades of railway automation. Empirical evidence on workload, fatigue, and human failure performance, obtained in multiple laboratory studies with professional train drivers participating in simulated work at three different grades of railway automation, was presented. In line with the hypotheses, results showed detrimental effects of intermediate grades of railway automation in combination with increasing monotony on workload, fatigue, and human failure performance. Further results showed beneficial effects of the domain‐specific decrease in monotony on workload, fatigue, and human failure performance at high grades of railway automation featuring remote control and recovery of unattended railway operation. The presented results stress the impact of domain‐ specific contextual factors, such as monotony in railway automation, on automation effects. Task characteristics were shown to play a major role in shaping workload, fatigue, and human failure performance. The influence of contextual factors on the generalizability of established effects of degrees of automation on workload, fatigue and human failure performance is discussed.