Data-Driven Detection and Assessment for Urban Railway Transit Driver Fatigue in Real Work Conditions
Bibliographic record
Abstract
Fatigue among urban railway transit (URT) drivers affects their performance and is a contributing factor in many railway accidents and incidents. This paper attempts to develop a robust fatigue detection system for URT drivers. An experimental study was conducted in actual work conditions, involving 198 professional URT drivers, to provide authentic and representative data. Fatigue scores based on the Karolinska Sleepiness Scale were used as the ground truth, and heart rate variability (HRV) data were collected using wearable photoplethysmography (PPG) sensors under actual working conditions. An extensive statistical analysis found that continuous working hours were a major factor in driver fatigue. HRV features were able to differentiate various fatigue levels. Four classifiers (k-nearest neighbors, Naive Bayes, support vector machines, and random forests) were trained to detect fatigue in real time for binary and three-class fatigue classifications, respectively. For the binary classification, the best performance was achieved by the random forest classifier using the corrected feature set as input with an accuracy of 92.5%. However, the accuracy dropped by 8 to 27 percentage points for the three-class classification. Moreover, the research found that the corrected feature set circumventing inter-individual variability in HRV could improve the performance of fatigue classifiers. The findings from this research could contribute to developing a robust and real-time URT driver fatigue detection system and improve current URT operational safety regulations.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".