Performance Prediction Model on Driving Train Simulator Based on Alertness and Sleepiness Level Result

Ayu Pramiarsih; Siswanto, Daniel; Susanto, Sani · 2021 · Crossref

DOI: 10.46923/ijets.v3i2.131

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the critical safety issue of train accidents caused by driver fatigue, specifically focusing on how sleep duration and sleep quality impact driving performance. Motivated by the high incidence of human-error-related accidents in Indonesia, the research aims to develop a predictive model for train driving performance based on objective and subjective measures of alertness and sleepiness. The study posits that fatigue, resulting from insufficient sleep or poor sleep efficiency, reduces alertness and increases the likelihood of operational errors, such as speeding. The experimental design employed a within-subject method involving 12 participants (8 for model building, 4 for validation), aged 22–38. Participants underwent four distinct sleep treatments: 4 hours or 8 hours of sleep duration, combined with good or poor sleep quality (defined by sleep efficiency ≥85% or <85%, respectively). Sleep data were recorded using Fitbit Charge 2 devices. Following each sleep condition, participants completed a 120-minute train simulation using "Train Simulator 2016." Driving performance was quantified by the percentage of time spent speeding. Alertness and sleepiness were measured using objective tools, including the Psychomotor Vigilance Test (PVT) and Sustained Attention Test (SAT), and subjective tools, including the Karolinska Sleepiness Scale (KSS), Swedish Occupational Fatigue Inventory (SOFI), and Visual Analog Scale (VAS). The results indicated that both sleep duration and sleep quality significantly affected driving performance, as confirmed by Repeated Measures ANOVA. Specifically, shorter sleep durations and poorer sleep quality correlated with higher percentages of speeding. The researchers constructed three multiple linear regression models to predict performance: one using only objective measures, one using only subjective measures, and one combining both. The combined model yielded the highest adjusted coefficient of determination ($R^2$) of 61.2%, indicating that 61.2% of the variance in driving performance (% speeding) was explained by the model. The final predictive variables included Mean Reaction Time from the PVT (MRTPVT), Percentage of Missed Targets from the SAT (NMTSAT), and Physical Exertion from the SOFI (PE). The model was validated using a Paired-T test on the separate validation group, confirming its statistical validity. The significance of this research lies in the development of a practical tool for assessing driver readiness. By integrating objective cognitive tests and subjective fatigue reports, the model can predict a driver’s likelihood of performance degradation before they begin their shift. This allows for proactive safety interventions, such as adjusting sleep schedules or substituting drivers, thereby mitigating the risk of accidents caused by fatigue-induced alertness deficits. The study concludes that monitoring both sleep hygiene and real-time alertness metrics is essential for enhancing railway safety.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• drowsiness detection algorithms
• sleep deprivation
• drowsiness
• time on task
• truck driver fatigue
• circadian factors

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: physiological data
• Methodological Resource: validation psychometrics
• Theoretical Contribution: computational model