Multi-state ship traffic flow analysis using data-driven method and visibility graph

Abstract

Ship traffic flow characteristics play a crucial role in enhancing the effectiveness and efficiency of intelligent maritime traffic management systems. The primary objective of this study is to establish a comprehensive framework for analyzing multi-state traffic flow based on the automatic identification system (AIS). The collected AIS data undergoes preprocessing to calculate traffic flow density, velocity, and intensity. Subsequently, clustering techniques, specifically the K-medoids algorithm and silhouette coefficient analysis, are applied to classify traffic states ranging from least congested to highly congested. The datasets corresponding to each cluster are then utilized to construct visibility graphs, which enable a graphical representation of the traffic flow dynamics. Statistical analysis is conducted to examine the topological characteristics of the network. To illustrate the applicability of the proposed framework, a case study of the Meishan island water areas is conducted, allowing for an in-depth analysis of ship traffic flow characteristics and the identification of distinct traffic flow states. The findings of this study demonstrate the effectiveness of the visibility graph method in analyzing multi-state ship traffic flow. Additionally, the statistical characteristics derived from the developed complex networks adeptly capture the inherent maritime traffic flow characteristics. The insights gained from this study contribute to the advancement of maritime traffic management by providing a deeper understanding of complex traffic flow patterns and delineation.