A real-time ship collision risk perception model derived from domain-based approach parameters

Artikel i vetenskaplig tidskrift, 2022

Ship collision is one of the most important factors that affect navigation safety, to perceive potential ship collision risk in advance is one of the effective means to reduce collision accidents. This paper proposes a real-time ship collision risk perception model which is derived from two domain-based approach parameters, DDV (Degree of Domain Violation) and TDV (Time to Domain Violation). This model considers the uncertainty of ship position prediction, adopts a novel reciprocal calculation method, and integrates the route information of own ship which is compliant with sailing practice during the parameter calculation process, then a collision risk identification method is formed, the roles of this model played in “manual loop” and “unmanned loop” decision-making processes are discussed respectively. Finally, the model is verified by two case studies, including a simulation case with nine target ships and a real case obtained from Automatic Identification System (AIS) data, several comparative analyses are conducted which demonstrate the proposed model's usefulness.