Future Rail Signaling: Cyber and Energy Resilience Through AI Interoperability

In today’s world, everything changes at lightning speed, making what is relevant today potentially obsolete tomorrow. This author’s scientific article addresses the issues of energy resilience and cybersecurity in railway signaling. A new proposal based on artificial intelligence is made to improve the fault tolerance of rail transport signaling infrastructure by ensuring increased energy efficiency and detecting cyber-attacks in real time. A linearly coupled neural network model was designed and implemented in a railway signaling simulation to simultaneously track the energy characteristics of signaling and detect abnormal behavior. The authors’ model was validated based on MATLAB(24.2.0.2863752 (R2024b) Update 5) simulations of a real double-track railway line under normal operating conditions and in a ransomware cyber-attack scenario. The AI simulation model correctly predicted the resilience of the signaling system, achieving an average absolute error of 0.0331 in predicting the fundamental performance indicator, and successfully identified an upcoming cyber-attack 20 min before the incident. This study demonstrates the promising architecture of the AI-based signaling system, which provides a significant increase in resilience to emergency situations in relation to power supply and cyber-attacks. By optimizing the signaling infrastructure with AI, it is possible to ensure safe and continuous movement of trains, including emergency situations, representing a promising approach to improving the resilience and safety of railways.