Effectiveness of Cognitive-Based and Conventional Alerters for Locomotive Engineers

Abstract

This is the first phase of a three-phase study that examined the effectiveness and reliability of alertness monitoring and warning devices for locomotive engineers. Four university students participated in this initial phase. Two experimental sessions were run for each participant, once during dayshift hours (9am to either 1pm or 5pm) and once during nightshift hours (10pm until 2am or 6am). They completed a series of vigilance performance tasks throughout their 4-hour or 8-hour experimental session. A conventional electro-mechanical alerter commonly found on board a locomotive was used for one of the two sessions, while a novel simple cognitive-based task was used for the other session. The Psychomotor Vigilance Test (PVT) and percent eye closure (PERCLOS) were used as measures of alertness. As expected, PERCLOS and performance task reaction time (RT) increased throughout the night sessions. Furthermore, it was found that when participants were exposed to the cognitive-based alerter, they had lower PERCLOS values throughout the night sessions and shorter vigilance task RT throughout the day sessions than when they were exposed to the conventional alerter. These results are discussed with respect to adaptive automation and their potential implications.