With the global advancement of maritime autonomous surface ships (MASS), the critical task of verifying their key technologies, particularly in challenging conditions, becomes paramount. This study introduces a synthetic maritime traffic generation system (S-MTGS) designed for the efficient and safe verification of these technologies. The S-MTGS encompasses a maritime traffic generator integrating a generator based on absolute position (GAP) and a generator based on relative position (GRP). This innovative system leverages historical maritime data to create various scenarios or generate virtual ships based on their interactions with and proximity to other ships. The virtual ships adeptly navigate and perform collision avoidance maneuvers with nearby vessels enabled by the integrated collision avoidance algorithm. The S-MTGS’s ability to generate a wide range of maritime traffic information mirroring actual maritime conditions is pivotal for thoroughly verifying the performance of MASS technology under both standard and extreme situations. The development of the S-MTGS represents a significant advancement in maritime safety and technology. It can evaluate collision avoidance and navigation systems in MASS, featuring a virtual environment for realistic scenario testing and an intelligent navigation system focused on route tracking and collision avoidance.

Although voice communication plays a crucial role in maritime operations due to its safety and efficiency, Maritime Autonomous Surface Ships (MASS) operating unmanned have not been fully integrated with existing voice communication systems. Therefore, this study proposes an integrated voice communication system for MASS. The proposed system combines state-of-the-art Speech-to-Text (STT) technology, Text-to-Speech (TTS) technology, and Large Language Models (LLM) to enable MASS to autonomously interact via VHF voice communication. In particular, the LLM generates appropriate responses that comply with standardized maritime communication protocols by learning Standard Marine Communication Phrases (SMCP) using Retrieval-Augmented Generation (RAG). The developed system enhances the autonomy of MASS and enables seamless interaction with existing manned ships and control centers. Experimental results demonstrated high speech recognition accuracy and real-time processing performance, confirming its practicality in maritime communication environments.

The dynamic behavior of autonomous surface ships is significantly influenced by environmental factors such as wind and currents, which introduce trajectory divergence and disrupt navigation stability. These disturbances pose significant challenges to maintaining accurate speed over ground (SOG) and course over ground (COG), particularly in real-time maritime operations. Traditional control methods, such as Proportional-Integral-Derivative (PID) controllers, often struggle to adapt to such disturbances, leading to suboptimal corrections and instability in navigation. To address these challenges, we propose a hybrid framework combining Fuzzy Logic Control (FLC) and an Extended Kalman Filter (EKF). The FLC provides real-time adaptability to varying environmental conditions, dynamically correcting SOG and COG based on divergence data. These corrections are integrated into the EKF, which leverages the improved state variables to enhance trajectory prediction accuracy. This dual-layer approach ensures robust control and accurate trajectory alignment even in challenging maritime environments. Comparative evaluations demonstrate the superior performance of the proposed framework. For control accuracy, FLC achieved RMSE values of 1.6648 knots and 28.3427 degrees for SOG and COG, respectively, compared to PID's 5.9102 knots and 55.0976 degrees. Similarly, FLC reduced MAE to 1.3212 knots and 14.4058 degrees for SOG and COG,