A New Risk Assessment Method for Surface Vessels

Abstract

This paper proposed a new risk assessment algorithm for surface vessels. The algorithm generalizes the qualitative risk assessment of ship domain method by introduce the support vector machine (SVM) algorithm to calculate the distance between ship domains such that the collision risk can be evaluated quantitatively. Two different cases are considered to assess the risk of own ship and target ship. In the case that the ship domains are separated, the maximum margin is used to measure the risk. In the case that the ship domains have intersected, the soft-margin defined by SVM to classify the nonseparable data is introduced to measure the risk. A framework like a classifier generated by SVM method, which combines the hard-margin and soft-margin of ship domains, can quantify the risk of ship domains. Simulation example is provided to illustrate the effectiveness of the proposed algorithm.