Probabilistic assessment of vessel collision risk: An evidential reasoning-based methodology

Feng Ma, Yu-Wang Chen

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review


This chapter proposes a novel method to estimate the collision probabilities of monitoring targets for coastal radar surveillance. Initially, the probability of a monitoring target being a real moving vessel is estimated using the records of manual operations and the Evidential Reasoning (ER) rule. Subsequently, the bridges, piers and other encountering vessels in a waterway are characterized as collision potential fields using an Artificial Potential Field (APF) model, and the corresponding coefficients can be trained in terms of the historical vessel distributions. As a result, the positional collision potential of any monitoring vessel can be obtained through overlapping all the collision potential fields together. The probabilities of authenticity and the collision potential are further formulated as two pieces of evidence on which the Dempster’s rule of combination is used to reason the collision probability of a monitoring target. The vessels associated with high collision probabilities can be highlighted for supervisors’ attention, as they potentially pose high risks to safety. A preliminary field test was conducted to validate the proposed method.
Original languageEnglish
Title of host publicationMulti-Criteria Decision Making in Maritime Studies and Logistics
Subtitle of host publicationApplications and Cases
EditorsPaul Tae-Woo Lee, Zaili Yang
PublisherSpringer London
Number of pages27
Publication statusPublished - 25 Oct 2017


Dive into the research topics of 'Probabilistic assessment of vessel collision risk: An evidential reasoning-based methodology'. Together they form a unique fingerprint.

Cite this