Complex event processing for network anomaly detection in digital railway communication services

The paper aims to propose a novel approach to rise the situation awareness about incidents, which can be considered as events in the complex event processing system. The central part of the proposed approach is network anomaly detection method based on fast harmonic analysis technique of telecommunication traffic and decision-making about abnormality in system functioning. The novel presented approach, as a whole, develops a new method that conforms main requirements of a brand new class digital railway network services, such as large data traffic volumes and real-time situation awareness. The developed methodology allows artificial intelligence algorithms of incident detection being embedded in the existed multilevel intelligent control system in Russian railway transportation.