Predict Maneuvering Indices Using AIS Data by Ridge Regression

Abstract

Since Automatic Identification Systems (AIS) was widely used in navigation industry, it has not only greatly benefited ship operations to keep encountered vessels visible to each other, but also fosters the relevant researches. Because AIS message contain vessel's identity, position, course, speed and so on, the application of AIS data make it possible to investigate accurate and actual behaviors of ships during ship maneuvering. The ridge regression is presented for identify maneuvering indices in Nomoto’s model, and the result indicates that the method is robust and does not rely on initial estimation. For selecting appropriate AIS data for maneuvering indices predicting, a frequency domain identification method is presented. Integrated with the predicted indices, the ship maneuvering mathematical model can be used to simulate the vessel traffic flow and assess the risk of collision.