STSRS:A New Dataset for Simulating Security Threats in Smart Railway Systems

The digitalisation of railway systems has improved efficiency but introduced vulnerabilities in train-to-control communications. To address the lack of realistic datasets for cyber-physical railway systems, this paper introduces STSRS (Security Threats in Smart Railway Systems), a high-fidelity time-series corpus for cybersecurity and intelligent transport research. STSRS contains over 10 million synchronised records simulating interactions between 10 autonomous trains and two control centres under both normal and adversarial scenarios (DoS, Jamming, Stealthy Replay). Generated using a modular framework with stochastic dynamics, TCP/IP telemetry, and context-aware attack injection, each record includes over 10 labelled features spanning raw (e.g., speed, latency) and behavioural (e.g., Burstiness, Entropy) metrics. STSRS supports reproducible anomaly detection, cyber-defence modelling, and digital twin evaluation in railway CPS environments.