PERANCANGAN APLIKASI DETEKSI KELELAHAN BERDASARKAN KECEPATAN REAKSI UNTUK PENGUKURAN FITNESS FOR DUTY PADA MASINIS DI PT KERETA API INDONESIA

• Main Author: Istri Arynda Pradnya Suary, A.A
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/87521
• Institution: Institut Teknologi Bandung

Fatigue and drowsiness are among the factors that cause accidents in various transportation sectors. The monotonous but high-risk nature of train driving requires high alertness while driving. Data published by the National Transportation Safety Committee (NTSC) shows that in Indonesia, 80% of train accidents are caused by human error and 45% involve driver fatigue. This fatigue is common in shift work systems and limits the sleep duration of machinists. Currently, the evaluation of alertness is done through medical examinations and interviews, but these methods are not accurate enough. Therefore, a valid, accurate, and sensitive vigilance test is needed. The Psychomotor Vigilance Task (PVT) is recognized as the gold standard in measuring vigilance through reaction speed. However, until now, no study has evaluated the PVT's performance in the context of machinist work and established its cut-off value as a reference. This research will use PVT as an objective measurement tool and the Karolinska Sleepiness Scale (KSS) questionnaire as a subjective measurement tool of fatigue to answer the formulation of research problems, namely determining the effect of workload driving a train simulation for one hour on reaction speed, the effect of wakefulness for 24 hours on reaction speed, and determining the cut-off value for the development of Fatigue Detection Technology applications. The analysis used several statistical methods, such as the Saphiro-Wilk test, Wilcoxon Signed-Rank test, Spearman Correlation test, Binary Logistic Regression test, and Receiver Operating Characteristic (ROC). The results of the analysis showed that a monotonous workload driving a train simulator affected decreasing reaction speed. Likewise, being awake for 24 hours can slow reaction speed due to the effects of fatigue or drowsiness. The results of PVT measurements that can be used as a reference for vigilance evaluation are medianRT with a cut-off value of 241 ms which will then be used as input for the development of Fatigue Detection Applications. This application is expected to be able to measure fitness for duty for machinists before duty and can contribute to improving the safety of railway transportation.