Decision support for preventing safety violations

Aim. The aim of the paper is to examine the experience of reducing the effect of the human factor on business processes, to develop the structure and software of the decisionsupport system for preventing safety violations by train drivers using machine learning and to analyse the findings. Methods. The study presented in the paper uses machine learning, statistical analysis and expert analysis. In terms of machine learning, the following methods were used: logistical regression, random forests, gradient boosting over decision trees with frequency-domain representation of categorical features, neural networks. Results. A set of indicators characterizing a train driver’s operation were identified and are to be used as part of the system under development. The term “train driver’s reliability” was defined as the ability not to violate train traffic safety over a certain number of trips. Algorithms were designed and examined for predicting violations in a train driver’s operation that are used in defining reliability groups and lists of preventive measures recommended for the reduction of the number of safety violations in a train driver’s operation. Major violations with proven guilt of the driver that may be committed within the following 3, 7, 10, 20, 30, 60 days were chosen as attributes for the purpose of safety violation prediction. Analysis of the results on the test sample revealed that the model based on gradient boosting over decision trees with frequency-domain representation of categorical features shows the best results for binary classification on the prediction horizon of 30 and 60 days. The developed algorithm made a correct prediction in 76% of cases with the threshold value of 0.7 and horizon of 30 days and in 82% of cases with the threshold value of 0.9 and horizon of 60 days. The solution of the problem can be found in the integration of different approaches to predicting safety violations in a train driver’s operation. Additionally, 10 of the most significant indicators of a train driver’s operation were identified with the best of the considered models, i.e., gradient boosting over decision trees with frequency-domain representation of categorical features. Conclusion. The paper presents an overview of methods and systems of assessing human reliability and the effect of the human factor on the safety of transportation systems. It allowed choosing the most promising directions and methods of predictive analysis of a train driver’s operation, including methods of machine learning. The resulting set of indicators of a train driver’s operation that take into consideration the changes in the quality of such operation allowed obtaining initial data for training the models implemented as part of the system under development. The implemented models enabled the aggregation of information on train drivers and adoption of targeted and temporary preventive measures recommended for improving driver reliability. The resulting approach to the definition of preventive measures has been implemented in three depots of JSC RZD in trial operation mode.

Robust Tactical Crew Scheduling Under Uncertain Demand

We analyze a tactical freight railway crew scheduling problem, when train drivers must be informed several weeks before operations about the start and end times and locations of their duties. Between informing the train drivers and start of operations, trip demand changes due to cancellations, new bookings, and reroutings of trains, which might result in mismatches between train driver capacity at a location and demand. We analyze an approach that incorporates uncertain trip demand as scenarios, such that the start and end times and locations of the duties of a crew schedule are recoverable robust against deviations in trip demand. We develop a column generation solution method that dynamically aggregates trips to duties and decomposes the subproblems into smaller, computationally tractable instances. Our model determines duty frames that cover duties in many scenarios, creating recoverable robust crew schedules. We test our model on three real data sets of a major European freight railway operator. Our results show that our schedules are considerably more recoverable robust than those of the nominal solution, resulting in smaller mismatches between train driver capacity and demand.

STRESS AND MONOTONY AS FACTORS OF LOCOMOTIVE CREWS’ SHIFT WORK

The article presents the results of a psychophysiological examination of drivers of passenger and export traffic, working in the day and night shifts, respectively. The dynamics of performing psychophysiological tests (changes in the reaction rate, the number of errors, perception of time intervals, etc.) after day and night shifts was revealed, and differences in the subjective perception of the specifics of work in the daytime and at night were also noted. Working the night shift requires the driver to mobilize psychophysiological resources aimed at maintaining active wakefulness and fighting monotony. Day trips are perceived to be more stressful due to more input and traffic. Target: The study of the functional state dynamics of train drivers working without an assistant during day and night trips with an increase in the duration of working hours up to 8 hours in passenger traffic and up to 12 - in export traffic. Methods: 1. Express test of the functional state; 2. «Sense of time» test; 3. Stress resistance test; 4. Survey «Well-being. Activity, mood (SAN)»; 5. Survey «Diagnostics of states of reduced performance (DORS)». Results: The dynamics of the speed and stability of the visual-motor reaction, the accuracy of the perception of time intervals, as well as the subjective perception of the features of day and night shifts by train drivers themselves, makes it possible to distinguish differences in the specifics of shift work: in the daytime it is distinguished by greater intensity, tension, which is reflected in the number of erroneous actions during testing after a day's ride on the simulator, and in the subjective experience of stress noted by the drivers. The need to work at night requires considerable efforts from train drivers to mobilize, which is manifested during a psychophysiological examination before the night shift, however, forced wakefulness during night work leads to a state of monotony among train drivers.

QUESTIONS OF PROFESSIONAL SUITABILITY AND REHABILITATION OF PERSONS AFTER CANCER

One of the features of the work of railway transport is the continuity of the labor process and ensuring the safety of this process. Even in modern technological conditions, there are deviations from hygienic standards at workplaces, which can lead to the development of various diseases, including cancer. When a tumor process is detected, issues related to special treatment are initially resolved. The issues of returning to the profession can become when the process is stabilized. A properly conducted rehabilitation process should be continuous and allow the patient to return to work and socially adapt. The incidence of malignant neoplasms on the West Siberian Railway is 1.5-1.9 times lower than the national indicators. At the same time, tumor processes in women are detected already at the age of 20-24 years, and in men 10 years later. When analyzing oncological morbidity, depending on the length of service, 2 peak indicators are observed. These are groups with experience of 11-15 years and 25-30 years. There is a high incidence of malignant neoplasms in train drivers, train drivers' assistants and track fitters. However, the group of social workers also showed high levels that can be associated with the influence of stress and the need to work at night.

The Impact of a Dynamic Working Environment on Human Gut Microbiota

Gut microbiota dysbiosis is associated with a variety of diseases, such as inflammatory bowel disease (IBD) and irritable bowel syndrome (IBS), metabolic diseases, allergic diseases, neurodevelopmental disorders and cancer. The human gut microbiota can be influenced by a variety of factors, including geography, dietary habits, living environment, age and altered lifestyle etc. This study was conducted to explore the gut microbiota compositions in officials who are in a stable working environment and train drivers who are in a dynamic working environment. Microbiota communities in the feces of 80 officials and 88 train drivers were analyzed using Illumina MiSeq sequencing targeting the V3-V4 region of 16S ribosomal RNA (rRNA) gene and ITS1 region of fungi. There were significant differences between the two groups in diversity and richness of gut microbiota, while the microbial community compositions of the two groups were similar. The relationship between gut microbiota and clinical characteristics was investigated. We found that more bacteria and fungi were positively correlated with clinical characteristics. Functional prediction analysis of the gut microbiota between the two groups by PICRUSt2 revealed significant differences between the official group and the train driver group. Elucidating these differences of the microbiome between the two groups will provide a foundation understanding of the impact of a dynamic environment on gut microbiota.

Electrophysiological Brain-Cardiac Coupling in Train Drivers during Monotonous Driving

Electrophysiological research has previously investigated monotony and the cardiac health of drivers independently; however, few studies have explored the association between the two. As such the present study aimed to examine the impact of monotonous train driving (indicated by electroencephalogram (EEG) activity) on an individual’s cardiac health as measured by heart rate variability (HRV). Sixty-three train drivers participated in the present study, and were required to complete a monotonous train driver simulator task. During this task, a 32 lead EEG and a three-lead electrocardiogram were recorded from each participant. In the present analysis, the low (LF) and high frequency (HF) HRV parameters were associated with delta (p < 0.05), beta (p = 0.03) and gamma (p < 0.001) frequency EEG variables. Further, total HRV was associated with gamma activity, while sympathovagal balance (i.e., LF:HF ratio) was best associated fronto-temporal delta activity (p = 0.02). HRV and EEG parameters appear to be coupled, with the parameters of the delta and gamma EEG frequency bands potentially being the most important to this coupling. These relationships provide insight into the impact of a monotonous task on the cardiac health of train drivers, and may also be indicative of strategies employed to combat fatigue or engage with the driving task.