Improved Z-number based Bayesian network modelling to predict cyber-attack risk for maritime autonomous surface ship (MASS)

Maritime Autonomous Surface Ship (MASS) will represent a transformative shift in the maritime industry, promising enhanced efficiency, and sustainability in transportation. It has advanced technologies, including artificial intelligence and sensor systems, to sail at sea without ship crew on board. Developed technology may face cyber threats as in many other areas. This situation may jeopardise not only operational continuity but also international maritime safety and the sustainability of global trade. Although some significant studies have explored cyber risks in ship navigation systems, they often lack a detailed probabilistic risk assessment tailored to fully autonomous vessels. Unlike previous studies, this research conducts a comprehensive probabilistic risk assessment focusing explicitly on cyber threats during the navigation of fully autonomous vessels. To do this, robust modelling including the Bayesian network (BN) is adopted under the improved Z-numbers theory. In the modelling, the BN is a significant tool capable of representing the cause-and-effect network between the variables, while the improved Z-numbers enable tackling uncertainties and enhancing the reliability of expert judgments. The findings of the research reveal that the occurrence probability of cyber-attack risk for MASS (degree 4- fully autonomous ship) is 7.15E-02 during navigation at open sea. Besides its robust theoretical background, the outcomes of research provide significant contributions to potential autonomous ship operators, MASS operators, ship inspectors, designers, maritime regulatory bodies and maritime security researchers for understanding and mitigating the potential cyber-attack threats and risk.