A predictive analytics method for maritime traffic flow complexity estimation in inland waterways. (April 2022)

• Record Type: Journal Article
• Authors: Zhang, Mingyang; Zhang, Di; Fu, Shanshan; Kujala, Pentti; Hirdaris, Spyros
• Is Part Of: Reliability engineering & system safety. Volume 220(2022)
• Publication Date: 2022-04
• Subjects: Traffic safety -- Maritime operations -- Inland waterways -- Artificial intelligence -- Lempel-Ziv algorithm -- TOPSIS; Reliability (Engineering) -- Periodicals; System safety -- Periodicals; Industrial safety -- Periodicals; Fiabilité -- Périodiques; Sécurité des systèmes -- Périodiques; Sécurité du travail -- Périodiques; 620.00452
• DOI: 10.1016/j.ress.2021.108317
• Languages: English
• ISSNs: 0951-8320

Abstract

Highlights: A predictive analytics method for maritime traffic complexity estimation is proposed. The method makes use of AIS to obtain ship travel time sequence. The nature of maritime traffic complexity is identified from microscopic perspective. Traffic complexity reflects the overall degree of irregularity and unpredictability. The method would be useful for ship traffic safety management. Abstract: Maritime traffic flow complexity is the factor that presents in most existing maritime safety analysis methods. It is considered as one of the main influencing factors affecting maritime safety. It can be estimated quantitatively through the analysis of traffic data. To explore maritime ship traffic in more detail, a predictive analytics method utilizing the Lempel-Ziv algorithm and the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) for traffic safety management is proposed. The Lempel-Ziv algorithm quantifies the entropy for the evaluation of the irregularity and unpredictability of ship travel time sequences, and TOPSIS ranks complexity. The results presented utilize Automatic Identification System (AIS) data corresponding to complex inland waterway traffic scenarios encountered in the Yangtze River. They show that high complexity implies ship travel time sequences are neither periodic nor stochastic, but dependent on the evolution patterns of traffic encounters. A correlation analysis of traffic flow complexity with the number of maritime accidents is carried out, showing that higher traffic flow complexity may result in more unwanted events. It is therefore concluded that the proposed method may help to (1) differentiate traffic flow complexity accurately, indicating that the higher the complexity value is, the higher the irregularity and unpredictability of maritime traffic flow are, (2) provide helpful reference for optimizing traffic management over the life cycle of fleet operations as well as maritime safety management of high traffic flow complexity areas.