Proactive Maritime Traffic Management Utilizing Dynamic Accident Networks (DANs)

Maritime safety in high-traffic areas, including narrow waterways, Traffic Separation Schemes (TSS), and port approaches, is a long-standing concern requiring advanced tools to address multifactorial risks and support operational decision making. Despite advancements in maritime technology and navigation systems, operational conditions, human and organizational factors remain critical in accident causation. Even in the era of autonomy and digitalization it is relevant, where manned, partially unmanned, and unmanned ships will coexist across different degrees of autonomy (D1-D4). Traditional accident analysis models often lack the capacity to dynamically incorporate real-time and complex operational variables. The maritime industry mostly relies on occurrence and casualty investigation reports to learn from the past and prevent in the future. To move towards proactivity, this study introduces a dynamic accident prediction model that integrates the Human Factors Analysis and Classification System (HFACS) with Bayesian Networks to assess and mitigate risks effectively in real-time and support operational decision making. Utilizing marine casualty data and incorporating expert insights, the model provides a probabilistic framework to support proactive decision-making for vessel traffic service (VTS) operators, ship masters and other maritime stakeholders. The model's application aims to enhance navigational safety and operational efficiency, by offering insights for decision making that will ultimately reduce accidents and enhance sustainable maritime operations.