Exploitation of maritime domain ontologies for anomaly detection and threat analysis

This paper describes a proof-of-concept prototype of an automated reasoning capability exploiting ontologies expressed in description logic to support the maritime staff in detecting anomalies, in classifying vessels of interest, and in identifying and categorizing maritime threats. The paper discusses the meticulous development of a maritime domain ontology of a significant size and the development of an automated reasoning service for the exploitation of this ontology. Practical exploitation examples are presented, making use of a description logic inference engine to detect anomalies and classify vessels of interest. Performance issues with the initial prototype are highlighted, and recent improvements are described that make the approach more suitable to match the high processing demands typical of realistic maritime environments.