Development of a Complex Event Recognition System for Maritime Activity Surveillance

Abstract

This thesis explores vessel behaviour and maritime patterns using AIS (Automatic Identification System) data, with a specific focus on tankers operating in North America between 2019 and 2022. Leveraging millions of AIS records alongside a complementary Fixtures dataset, we develop a multifaceted analytical framework designed to extract operational, behavioural, and market-related insights from vessel movement data. The methodology includes data preprocessing, feature engineering, dataset integration, and spatial filtering to analyse vessel speed, port performance, anchoring patterns, and chartering behaviour. Our work is motivated by the increasing demand for data driven decision making in the maritime industry, where real time and historical trajectory analysis plays a critical role in enhancing operational efficiency, environmental awareness, and regulatory compliance. By integrating spatio-temporal AIS information with commercial chartering data, this research bridges the operational and market dimensions of maritime activity, offering a holistic view of tanker operations. The analysis reveals clear regional and temporal trends in port utilization, congestion, and anchorage behaviour across the North American coastline. Results show that tanker waiting times and laycan durations vary significantly between ports and seasons, reflecting the combined effects of market dynamics, port infrastructure, and trade intensity. The integration of Fixtures with AIS data enabled the identification of direct links between commercial demand and operational performance, demonstrating that peaks in fixture activity are often associated with measurable increases in port congestion and anchorage utilization. The developed framework successfully identified major trade corridors, recurring anchorage zones, and systematic differences in port efficiency, contributing to a more complete understanding of tanker traffic and its economic context. Overall, this study provides both methodological and empirical contributions to maritime data science. It demonstrates how combining operational and commercial datasets can produce reliable, interpretable, and actionable insights for stakeholders. The findings not only highlight the operational realities of tanker deployment in one of the world’s most active maritime regions but also establish a replicable analytical foundation for future studies in global vessel behaviour, freight market assessment, and data-driven maritime decision support.