Automatic Identification of Ship Navigation Risk for Collision Accidents Using Uncertain Regression Model

With the growth of quantity, scale and speed of vessels, lots of maritime accidents especially collision accidents are posing the risk to marine transportation, its safety has attracted increasingly attention. To improve the safety and efficiency of navigation, this paper proposes an automatic identification algorithm to improve the awareness of overall navigation risk that is defined as the probability of collision accidents. The selected ship navigation risk factors causing collision accidents are calculated and integrated by using fuzzy reasoning and a novel method of an uncertain regression model. To identify high collision risk of coastal restricted water areas, calculations of bad weather, strong tidal currents, traffic congestion, and fishing boat operating areas are addressed and an uncertain regression model is supposed to helpfully combine the overall navigation risk under varying marine traffic and external conditions. Applications are given to validate its effectiveness.