Big Data–Based Estimation for Ship Safety Distance Distribution in Port Waters

Abstract

The water area of a large container port such as Singapore's has high ship traffic density because of continuously increasing international seaborne trade. The behavior of ships sailing in the port's waters exhibits high diversity. Ships must maintain a minimum safety distance when moving in and out of port waters to avoid collisions. This study estimated the probability distributions for ship safety distance by using automatic identification system (AIS) data collected in Singapore port waters. Thirty-six navigation scenarios classified by ship type and size, visibility (daytime and night), and direction of movement (crossing, head-on, and overtaking) were investigated. Safety distances for various ship types and sizes were first examined with nonparametric statistical tests. A tangible approach incorporating the maximum likelihood estimation and Kolmogorov–Smirnov test techniques was designed for determining the best-fitted probability distribution with the parameters calibrated by AIS data for ship safety distance. It was found that the lognormal and gamma distributions could well fit the ship safety distance in Singapore port waters according to the collected AIS data.