Towards a Systematic Safety Evaluation Framework for Mass in Congested Waters

Abstract A clear trend towards the maritime autonomous surface ship (MASS) has been observed within the maritime industry in recent years. Risk awareness and safety evaluation are proactive manners to advance safe and intelligent autonomy for the MASS system. This study presents a systematic safety evaluation framework for MASS in congested waters, aiming to enhance its safety performance while adhering to the International Regulations for Preventing Collisions at Sea (COLREGs). The proposed conceptual framework consists of three essential phases. Initially, a systems theoretic process analysis (STPA) method can be applied for risk identification from the control perspective for MASS, with the objective of evaluating its hazardous events, unsafe control actions, and potential causes. The outcomes of STPA are further incorporated into a Bayesian network (BN) to estimate the real-time risk level. In the final phase, the risk index and spatial-temporal risk map can be generated to facilitate risk-informed decision making. The proposed framework is validated through a case study that delivers a detailed analysis of using STPA-BN for safety evaluation towards unsafe control. It demonstrates how risk analysis methods can be applied and integrated with autonomous control systems, which could be further utilized to guide decision-making and adaptive behavior to enhance the degree of autonomy and system safety. The proposed framework in this study is tailored for the MASS system, but it also can be adapted to other marine autonomous systems.