A novel regional ship collision risk assessment framework for multi-ship encounters in complex waters

Abstract

To address the issue of precisely quantifying ship collision risk and identifying collision hotspots in complex waters, this study proposed a regional ship collision risk assessment framework that includes a multi-ship encounter recognition model, a multi-ship comprehensive collision risk assessment model, and a collision hotspot recognition model. Firstly, the directed distance density-based spatial clustering of applications with noise (DBSCAN) algorithm was proposed. Based on this algorithm, a multi-ship encounter recognition model was established to cluster ship positions and obtain ship encounter clusters. Subsequently, the multi-ship comprehensive collision risk assessment model was constructed for multi-ship encounter situations based on the hierarchical classification method, enabling the determination of each ship's Comprehensive Collision Risk Value (CCRV). Finally, the collision hotspot recognition model was built using the grid analysis method, which can identify areas with a higher risk of ship collision. The automatic identification system (AIS) data of the Yangtze River estuary were used to validate the effectiveness of our framework. The results show that our framework can accurately measure ship collisions risks and identify areas with high collision risks, offering valuable insights to maritime authorities, providing actionable insights for real-time ship collision risk monitoring.