Constructing a Decision-Support System for Safe Ship-Navigation Using a Bayesian Network

In a complicated sea environment, a ship needs experienced seafarers to steer it safely. Advancements in technology have not fundamentally reduced the number of accidents at sea. In deep-seated research and analysis of accident causes, human factors will always have an obvious or potential impact. Maritime transport is an arduous industry. Workers’ willingness to engage in seafaring occupations is gradually decreasing. This is an irreversible trend. The seafarer shortage and the seriousness of the safety situation are seemingly irreconcilable. To address this situation, the automation of ship equipment and intelligent decision-making must be accelerated. Autonomous decision-making is a basic step toward intelligent or unmanned navigation. The purpose of this paper is to construct a human-in-the-loop decision-support system for safe ship navigation, to minimize the impact of human factors, to reduce the accidents that occur because of poor human decision-making, and to ensure the ship will navigate safely at sea. This presupposes that a reliable decision-support system can be constructed. It requires relatively accurate predictions, based on past experience and objective accident-probability statistics. A Bayesian network can be used for risk and accident predictions. Therefore, the principles of a Bayesian network can be used for collision avoidance, and also for decisions on other sea conditions during a voyage. This paper discusses the prospects of an intelligent decision-support system to ensure reliable navigation safety, using a decision-support systematic approach, with a Bayesian network.