Maritime Anomaly Detection of Malicious Data Spoofing and Stealth Deviations from Nominal Route Exploiting Heterogeneous Sources of Information

Abstract

Based on a proper stochastic formulation of the vessel dynamic, exploiting piecewise Ornstein-Uhlenbeck (OU) mean-reverting processes, we propose an effective anomaly detection procedure to jointly reveal Automatic Identification System (AIS) data spoofing and/or surreptitious deviations from the planned route. Supported by reliable information from monitoring systems (coastal radars and spaceborne satellite sensors), an expanded five-hypothesis testing problem is posed involving two anomaly detection strategies based on the Generalized Likelihood Ratio Test (GLRT) and the Model Order Selection (MOS) methodologies.