JITA - APEIRON

Application of Artificial Intelligence Methods for the Prediction of Hazardous Failures

Igor Shubinsky, Alexey Ozerov — Wed, 03 Jan 2024 00:00:00 +0100

The availability of real-time data on the state of railway facilities and the state-of-the art technologies for data collection and analysis allow transition to the fourth generation maintenance. It is based on the prediction of the facility functional safety and dependability and the risk-oriented facility management. The article describes an approach to assessing the risks of hazardous facility failures using the latest digital data processing methods. The implementation of this approach will help set maintenance objectives and contribute to the efficient use of resources and the reduction of railway facility managers’ expenditures.

The Dependability and Safety Indicators of a Train Driver-Machine System

Igor Shubinsky, Efim Rozenberg, Natalia Boyarinova — Wed, 03 Jan 2024 00:00:00 +0100

The paper aims to assess the effect of the existing actions to assist a train driver in various operational situations, as well as to numerically evaluate the effect of such assistance on the resultant indicator of an error-free driver performance. The paper calculates and analyses the probability of at least one of the independent events or actions aimed at improving the quality of driver performance and reduction of the probability of error. The model of an environment was created, in which the probability of error-free driver performance is affected by a number of factors.

Evaluation of the Effect of Operational Scenarios on a Train Driver Performance

Igor Shubinsky, Efim Rozenberg, Natalia Boyarinova — Wed, 03 Jan 2024 00:00:00 +0100

This paper aims to numerically evaluate the effect of existing actions to assist a train driver in various operational situations, as well as select the indicators of error-free operation ending on the form of activity and other factors. The effect of each individual examined factor on the resulting indicator was evaluated, operational situations were examined taking into account the proportion of times when the action has a positive effect. A few practical cases were examined, whereas the method can be used.

Study of Training Quality of Multilayer Artificial Neural Networks with Variable Signal Conductivity in Scheduling Problems

Alexey Olshansky — Wed, 03 Jan 2024 00:00:00 +0100

The paper studies approaches to railway scheduling problems using artificial neural networks (ANNs). The authors analyze traditional learning algorithms and difficulties for their application. The description of ANN’s behavior is provided in the form of a phase portrait. New approaches and techniques are proposed for quality improvement of training of multilayer ANNs with variable signal conductivity.

Big Data-Based Methods for Functional Safety Case Preparation

Efim Rozenberg, Alexey Olshansky, Alexey Ozerov — Wed, 03 Jan 2024 00:00:00 +0100

The paper aims to overview the opportunities, approaches and techniques of studying and ensuring functional safety of transportation systems, including those driverless, with the use of Big Data. Examples are provided of machine learning/Big Data application in analysing the functional safety of complex control/management systems in railway transportation. The paper proposes the concept of application of supervised artificial neural networks combined with model checking. The following methods were used in the preparation of the paper: system analysis, logical and comparative analysis and historical principle. Updated requirements are defined for transportation systems using artificial intelligence as part of adaptive train schedule management and autonomous train control. That will ultimately allow developing an entire line of research from AI-based system functional safety estimation and machine learning to safety case preparation of intelligent supervised control/management systems based on formal verification.

Computer Vision as Part of an Advanced Train Control System

Alexey Ozerov, Andrey Okhotnikov — Wed, 03 Jan 2024 00:00:00 +0100

The article addresses the key issues related to the application of computer (machine) vision as part of an advanced train control system. The authors describe the architecture of a multi-layered train control system that uses computer vision as a key element of autonomy and automatic detection of obstacles on track ahead the train. The article provides some overview of computer vision sensors, key stages of dataset preparation for onboard perception, and some issues related to sensor calibration.

Cybersecurity Standards of Intellectual Transport Systems

Alexey Ozerov — Wed, 03 Jan 2024 00:00:00 +0100

The paper gives an overview of general approaches and existing standards as regards the cybersecurity of automated control systems with the railway transport specifics taken into account. It outlines major directions for the development of guidelines and activities for ensuring the cybersecurity of intellectual control systems.

Train Control Systems for High Speed Rails

Alexey Ozerov, Elena Denchik — Wed, 03 Jan 2024 00:00:00 +0100

The article provides an overview of train control and protection systems used on high-speed railways in given European and Asia-Pacific region countries. Particular attention is paid to currently operated and future means for the transmission of safety-related information to the trains.