Virtual Testing for MASS: A Comprehensive Review of Current Practices, Gaps, and Standardization Needs

Abstract

This paper presents a comprehensive literature review and landscape analysis of virtual testing for autonomous ship navigation and collision detection/collision avoidance (CDCA), with the objective of consolidating fragmented research and industrial practice into a structured overview that can inform future standards and certification frameworks for Maritime Autonomous Surface Ships (MASS). The review encompasses academic studies, regulatory and class-society publications, national and regional testbed activities, industrial developments and prototype testing initiatives, as well as early operational insights obtained from MASS simulators and sea trials. Methodologically, the study employs structured searches across scientific databases, regulatory documents, industrial white papers, trial reports, and vendor materials to map the current state of the art in testing of autonomous CDCA systems. The analysis covers the types of scenarios (baseline, region-specific, edge and near-miss), test procedures, performance metrics, pass/fail criteria, coverage requirements, existing IMO manoeuvre tests applicability and the practical approaches and validation strategies adopted by developers. The review highlights strong growth in scenario-based testing, digital twins, high-fidelity simulators, and human-in-the-loop arrangements, with increasing focus on safety and reliability indicators tailored to autonomous functions. At the same time, the review identifies critical gaps, including the limited standardization of safety metrics and acceptance criteria, inconsistent handling of critical/near miss or other edge scenarios derived from real traffic data, and challenges in the translation of virtual testing evidence into regulatory or class approval pathways. Drawing from these observations, the paper outlines a high-level concept of an autonomous CDCA virtual testing framework. The paper highlights overarching elements of the framework including consistent scenario development, clear evaluation metrics, harmonized safety and reliability indicators, and mechanisms for linking virtual testing outcomes to broader assurance pathways such as actual MASS trials, regulatory and class approvals, and crew training. The findings establish an important knowledge foundation and identify key areas from a maritime assurance perspective, motivating future efforts towards practical, standardized approaches to virtual testing of autonomous maritime systems.