Autonomous Ship Collision Avoidance Navigation Concepts, Technologies and Techniques

TLDR

Autonomous ship navigation must address factors influencing collision avoidance, and while human operators can navigate satisfactorily, their subjective decisions can lead to errors, prompting research that is divided into classical and soft‑computing approaches. The study aims to provide a comprehensive understanding of autonomous ship navigation for collision avoidance and its theoretical background. The authors examine human cognition, COLREGs, and water‑based collision avoidance algorithms, and review classical mathematical models and soft‑computing AI techniques—including evolutionary algorithms, fuzzy logic, expert systems, neural networks, and hybrid systems.

Abstract

This study provides both a spherical understanding about autonomous ship navigation for collision avoidance (CA) and a theoretical background of the reviewed work. Additionally, the human cognitive abilities and the collision avoidance regulations (COLREGs) for ship navigation are examined together with water based collision avoidance algorithms. The requirements for autonomous ship navigation are addressed in conjunction with the factors influencing ship collision avoidance. Humans are able to appreciate these factors and also perform ship navigation at a satisfactory level, but their critical decisions are highly subjective and can lead to error and potentially, to ship collision. The research for autonomous ship navigation may be grouped into the classical and soft computing based categories. Classical techniques are based on mathematical models and algorithms while soft-computing techniques are based on Artificial Intelligence (AI). The areas of AI for autonomous ship collision avoidance are examined in this paper are evolutionary algorithms, fuzzy logic, expert systems, and neural networks (NN), as well as a combination of them (hybrid system).