A study on Risk Analysis of Ship Collision Avoidance Maneuvers

Abstract

The aim of this graduation thesis is to propose a method to evaluate collision risk along ship trajectories, in order to assess the collision risk probability of different encounters. Navigating a vessel includes following a pre-planned trajectory while avoiding collisions as well as other possible risks and thus, accurate evaluations of risks are vital for safe navigation. Utilizing concepts in the domain of collision avoidance and risk assessment is, therefore, imperative in ship manoeuvres in both manned and autonomous/unmanned scenarios. Identifying collision risks along a particular trajectory not only aids in safe navigation but also is vital to assess navigators’ workload in respective path planning. Experimental simulations done in this study focuses on modelling ship encounters, which essentially avoid collision based on velocity obstacle algorithm, following International Regulations for Preventing Collisions at Sea (COLREGs), and assessing collision risk in different encounter scenarios. A simulation model was developed utilizing velocity obstacle algorithm for collision avoidance, molecular collision theory for the proposed approach for collision risk assessment, and conventional closest point of approach (CPA) model for validating collision risks in respective encounters. The simulation program was able to model head-on, crossing and overtaking scenarios between a particular vessel and multiple target-ships. Collision avoidance manoeuvres were successfully carried out, and collision risk was assessed following the two aforementioned approaches. Results obtained from the study supported the accuracy of the proposed model, thus validating the approach.