(03)

Science

↓ (W19 ‘26 · May 4–10)
01crossref2026-05-07L.P. PereraDOI ↗ (opens in a new tab)

Sandbox environment for evaluating safety-critical future autonomous ship navigation (opens in a new tab)

This paper examines the limitations of current autopilot systems in autonomous ships, which mainly handle basic navigation and lack strong collision avoidance. It proposes using VR simulators and model-scale vessel testing as a controlled "sandbox" environment to improve autonomous navigation. The study highlights the role of advanced predictors and AI in detecting and preventing close-encounter situations early. It also stresses the need to ensure both hardware and software reliability in AI systems to handle potential failures and maintain safety.

02openalex2026-05-07Manuel Alba Fernández, Juan Pablo Rodríguez DelgadoDOI ↗ (opens in a new tab)

Liability for damage caused by autonomous ships in light of the EU framework on artificial intelligence (opens in a new tab)

This paper examines the legal challenges posed by autonomous ships under EU AI liability frameworks. It highlights the shift from strict liability in maritime conventions to a fault-based approach in the EU’s 2022 AI Liability Directive and 2024 AI Act, which aims to ease victims' evidentiary burdens. The authors argue that preserving the shipowner’s central role in liability rules, as in traditional maritime law, ensures fair risk distribution and effective compensation. The study contrasts EU developments with existing maritime conventions like CLC 1992 and Bunkers 2001.

03openalex2026-05-07Henrik RingbomDOI ↗ (opens in a new tab)

Regulatory challenges linked to MASS (opens in a new tab)

The paper examines regulatory challenges posed by Maritime Autonomous Surface Ships (MASS) at the international level. Key issues include adapting MASS to the law of the sea, particularly regarding passage rights and flag state responsibilities. Technical regulation is progressing through the draft MASS Code, though gaps remain in areas like collision avoidance (COLREGs) and watchkeeping. The shift from human-operated to autonomous vessels also raises unresolved questions about civil liability for damages.

04openalex2026-05-07E. Anastopoulou, V. Karystinos, G.P. Trachanas, D.E. Liarokapis +1 moreDOI ↗ (opens in a new tab)

Ship model telemetry system and RL speed control system in simulated vessel (opens in a new tab)

This paper presents a real-time telemetry system for Maritime Autonomous Surface Ships (MASS) that collects sensor data via a Raspberry Pi 5, stores it in a PostgreSQL database, and visualizes it through a web application. It also develops an AI-based speed controller using the Soft Actor-Critic (SAC) reinforcement learning algorithm. The controller relies on an XGBoost model trained on Simulink vessel dynamics data to issue throttle commands. The system successfully tracks multiple speed profiles and serves as a foundation for autonomous marine operations.

05crossref2026-05-07M. Adnan, L.P. PereraDOI ↗ (opens in a new tab)

Transfer learning-based computing architecture to support autonomous maritime operations (opens in a new tab)

This paper explores challenges in Maritime Autonomous Surface Ships (MASS), including handling large sensor-generated data onboard and transferring it remotely. It highlights the reliance on AI for decision-making and the need for localized tuning of these systems. The authors propose a transfer learning-based computing architecture to improve autonomous maritime operations. A case study demonstrates its application in monitoring autonomous vessel performance.

06openalex2026-05-07Raphael EsuDOI ↗ (opens in a new tab)

Unmanned vessels (opens in a new tab)

The paper explores legal and security challenges posed by Maritime Autonomous Surface Ships (MASS) under the UK’s Aviation and Maritime Security Act (AMSA) 1990. It highlights that current frameworks, designed for physical threats, fail to address risks like cyberattacks and remote commandeering unique to autonomous vessels. The author proposes adapting AMSA to include a graded system for unauthorized interference, categorizing incidents by severity. This approach aims to strengthen maritime security and establish the UK as a leader in regulating autonomous shipping.

07openalex2026-05-07Eva Ricarda BaddenhausenDOI ↗ (opens in a new tab)

EU and MASS (opens in a new tab)

This paper analyzes the EU’s role in advancing autonomous vessels (MASS) for short-sea shipping, emphasizing its funding of research and integration into supply chains. It examines the EU’s regulatory authority in maritime transport and the ongoing IMO legislative process for MASS. The study compares European regulatory approaches and assesses their impact on Norway. The author recommends the EU adopt a balanced legislative strategy, aligning with international standards while ensuring legal clarity in its waters.

08openalex2026-05-07Barbara StępieńDOI ↗ (opens in a new tab)

Next-generation autonomous vessels and the IMO's decarbonization goals (opens in a new tab)

The paper examines how next-generation autonomous vessels may affect the International Maritime Organization’s (IMO) decarbonization goals. It analyzes environmental challenges and potential benefits of autonomous ships in reducing climate impact. The study highlights the need for IMO’s decarbonization strategies to align with technical evaluations and feasibility assessments. Adoption of environmental regulations for these vessels must be grounded in validated research.

09openalex2026-05-07D. Rhidian ThomasDOI ↗ (opens in a new tab)

Some thoughts on the insurance of autonomous ships (opens in a new tab)

The paper examines legal, regulatory, and insurance challenges tied to autonomous shipping, emphasizing the need for supportive frameworks. Key terms like "autonomous," "ship," and "cyber risk" are criticized for imprecision, complicating policy and coverage. Existing marine insurance, such as ITCH83 and P&I Club rules, may cover some cyber risks if losses are classified as marine, but gaps could require additional cyber-specific policies. Insurers are expected to demand strict cybersecurity practices as a condition for coverage.

10openalex2026-05-07Unho LeeDOI ↗ (opens in a new tab)

Navigating autonomy with human control (opens in a new tab)

The paper compares Japan and Korea’s regulatory approaches to maritime autonomous surface ships (MASS), focusing on human oversight and accountability. Japan uses evolving guidelines, like the 2025 MASS Safety Standards, to define roles for crews and remote operators in supervising autonomous systems. Korea’s 2023 MASS Act establishes a regulatory sandbox with exemptions for trials, requiring insurance and operational approvals. Both countries align with the IMO’s cautious, human-in-the-loop direction, emphasizing collaboration for safe global deployment.

11openalex2026-05-07E. Eliopoulou, K. Louzis, A. Koimtzoglou, D. Konovessis +1 moreDOI ↗ (opens in a new tab)

An integrated method for the early design of autonomous ships (opens in a new tab)

This paper introduces a structured method for designing autonomous ships by combining traditional ship design principles with autonomy-specific requirements. The approach helps generate early concept designs optimized for operational needs, assessing how autonomy affects ship dimensions, layouts, and systems. A case study on a feeder containership demonstrates the method’s application. Future research aims to expand the focus beyond the ship itself to entire autonomous maritime ecosystems.

12openalex2026-05-07Frank StevensDOI ↗ (opens in a new tab)

Autonomous ships – do we need a MASS master? (opens in a new tab)

The paper examines the evolving role of ship masters with the rise of remotely operated and autonomous vessels (MASS). Traditional authority of masters has already diminished due to improved communication technology and increased ship owner control. It questions whether appointing a master for MASS is legally necessary or the best solution. The author suggests exploring alternative approaches to address legal and operational challenges.

13openalex2026-05-07Zuzanna Pepłowska-Dąbrowska, Igor VioDOI ↗ (opens in a new tab)

Problems arising from the IMO's regulatory scoping exercises – paths to take (opens in a new tab)

The IMO’s Maritime Safety Committee (MSC), Facilitation Committee (FAL), and Legal Committee (LEG) conducted regulatory scoping exercises (RSEs) to evaluate existing conventions for maritime autonomous surface ships (MASS). A joint working group was formed in 2022 to address cross-cutting issues and propose amendments or new conventions. The MSC plans a non-mandatory safety code for MASS by 2026, followed by a mandatory code with SOLAS amendments by 2032. The Legal Committee will use non-binding guidelines to adapt relevant international instruments for MASS.

14openalex2026-05-07Melis ÖzdelDOI ↗ (opens in a new tab)

CMI and MASS (opens in a new tab)

The paper by Melis Özdel (2026) examines the role of the CMI International Working Group on Maritime Autonomous Surface Ships (MASS) in developing legal frameworks for autonomous vessels. It highlights the group’s growing influence in addressing challenges like liability, governance, and regulatory adaptation for remote-controlled and autonomous ships. The chapter critically discusses debates, including whether a strict liability regime should apply to MASS. It combines scholarly and practical insights to contribute to discussions on international maritime law evolution.

15openalex2026-05-07Mihael Mišo Mudrić, Kristijan KotarskiDOI ↗ (opens in a new tab)

AI, AI captain, MASS liability ahoy (opens in a new tab)

This paper examines how current and future regulations impact AI systems and product liability, specifically in the maritime sector. It compares existing transportation norms, particularly from the road sector, with maritime laws, focusing on AI integration. The analysis covers contractual and non-contractual liability rules under European legislation for defective AI products. The authors highlight the need for tailored AI regulations in maritime to address safety and liability, especially for autonomous ships.

16crossref2026-05-07S.H. Ahmed, L.P. Perera, M. AdnanDOI ↗ (opens in a new tab)

Mission planning framework for autonomous marine vehicles in seabed surveys (opens in a new tab)

This paper presents a mission-planning framework for autonomous marine vehicles conducting seabed surveys, focusing on safety, COLREG compliance, and operator transparency. It introduces a layered architecture for an onshore-supervised surface vessel that generates coverage plans from operator-defined polygons and produces structured reports for underwater vehicle inspections. The system includes a behavior-tree executive, a reinforcement-learning advisor for optimized survey edits, and a risk assessment layer for collision avoidance. The framework is validated through ongoing development of a surface vessel and remote operation center, using repeatable logs and status reporting.

17openalex2026-05-07Maria Pia BenosaDOI ↗ (opens in a new tab)

Laying the groundwork for MASS regulation at IMO (opens in a new tab)

The paper examines the International Maritime Organization’s (IMO) efforts to develop regulations for maritime autonomous surface ships (MASS). It reviews the early stages of the IMO’s rule-making process, highlighting key legal and technical challenges. The article identifies unresolved issues that must be addressed before adopting a formal legal framework for MASS. It focuses on the applicability of existing maritime laws to autonomous and remotely operated vessels.

18crossref2026-05-08Dui Liu, Hengyu Li, Xiang Liu, Yumeng Cao +2 moreDOI ↗ (opens in a new tab)

Resilient Path Tracking of Unmanned Surface Vehicles Under False Data Injection Attacks: An Adaptive Event-Triggered Approach (opens in a new tab)

This paper addresses cybersecurity threats to unmanned surface vehicles (USVs), focusing on false data injection (FDI) attacks that disrupt sensor data or communications. It introduces an adaptive control scheme with an adaptive threshold event-triggering mechanism (ATETM) to maintain stability and reduce communication demands. Novel coordinate transformations and the Nssbuam function help counteract time-varying attack effects. Simulations confirm the method achieves accurate path tracking while lowering control update frequency under deception attacks.

19openalex2026-05-04Luis F. W. Batista, Stéphanie Aravecchia, Cédric PradalierDOI ↗ (opens in a new tab)

Sim-to-Real Transfer and Robustness Evaluation of Reinforcement Learning Control with Integrated Perception on an ASV for Floating Waste Capture (opens in a new tab)

This study presents a field-tested system for floating-waste capture using an autonomous surface vessel (ASV) with camera-based perception and a deep reinforcement learning (DRL) controller. The system, trained entirely in simulation, was deployed on a real ASV, achieving centimeter-level accuracy in waste capture under various disturbance conditions. A two-stage simulation protocol and perception abstraction module helped evaluate the "sim-to-real" gap, revealing that actuation-model fidelity was the primary limitation. The approach was demonstrated in real-world search-and-capture tasks over areas up to 450 m², with lessons for improving reliability in sim-to-real transfer.

20crossref2026-04-13Shanthan K. Padisala, Satadru DeyDOI ↗ (opens in a new tab)

Model Predictive Control-Based Optimal Energy Management of Autonomous Electric Vehicles Under Cold Temperatures (opens in a new tab)

This paper presents a real-time Model Predictive Control (MPC) approach for optimizing energy management in autonomous electric vehicles (AEVs) under cold temperatures. It addresses the challenge of balancing battery power allocation between propulsion, HVAC systems, and battery preconditioning to preserve driving range and battery health. The method dynamically adjusts power distribution to maintain thermal comfort while ensuring the battery is ready for immediate charging upon arrival. The study aims to replace heuristic rules with a principled strategy for improved energy efficiency.

21crossref2026-06-01Hanna Hrudinina, Volodymyr Blintsov, Lech RowińskiDOI ↗ (opens in a new tab)

Automatic Course Stabilisation of an Autonomous High-Speed Boat Under Wind–Wave Disturbance Conditions (opens in a new tab)

This study developed a fuzzy logic-based controller to stabilize the course of a small, high-speed unmanned boat under wind and wave disturbances. The controller’s performance was tested via computer simulations, comparing its course-keeping accuracy and rudder activity to a traditional PID controller. Key evaluation metrics included course angle stability, variance, and rudder deflection frequency. The design prioritized energy efficiency by minimizing rudder movements while maintaining precision.

22crossref2026-05-06Yong Li, Dehang Lian, Jialong Du, Dongxu Gao +2 moreDOI ↗ (opens in a new tab)

Weak Calibration Cross-Fusion Framework for Multi-Modal 3D Object Detection on Unmanned Surface Vehicles (opens in a new tab)

This study introduces RCF-Free, a multi-modal 3D object detection framework for unmanned surface vehicles (USVs) that eliminates the need for precise sensor calibration. The method uses a Triple-Path Cross-View Fusion module, a Mobile Self-Attention Module, and a BEV-Point feature encoder to improve 3D perception in dynamic water environments. The authors also present two new water-related datasets, FloW-BEV and WaterScenes-BEV, achieving mAPBEV50 scores of 60.5% and 69.3%, respectively. The model demonstrates strong performance in both water and autonomous driving scenarios, with cross-domain generalization.

23openalex2026-05-05M. Preenu Antony, S. Sreepriya, Helga Joseph, S. JithendranDOI ↗ (opens in a new tab)

From Harbor to High Seas: A Qualitative Exploration of I-boat Awareness among Small-Scale Fishing Communities (opens in a new tab)

This study examines small-scale fishermen in Chellanam Harbour, Kerala, and their awareness of I-boat, an AI-driven autonomous vessel designed to improve navigation and ecosystem monitoring. Findings from interviews with 60 fishermen reveal limited and inconsistent knowledge of the technology. Key barriers include the lack of a Malayalam interface, inadequate infrastructure, and no formal training. The authors recommend localized, language-specific awareness programs led by the fisheries department to improve adoption.

24openalex2026-05-08Zhonglian Jiang, Di Yang, Zhen Yu, Changling He +1 moreDOI ↗ (opens in a new tab)

Causation Analysis of Maritime Accidents Based on Coupled AcciMap and GCN Model (opens in a new tab)

This study introduces a new model combining AcciMap and graph convolutional networks (GCN) to analyze causes of maritime accidents in the Yangtze River waterway. The model uses mutual information, a grey wolf optimizer, dropout layers, and weakly supervised learning to identify causal chains and rank accident causes by importance. It outperformed baseline models, achieving an MAE of 0.0820 and RMSE of 0.1694, with faster training convergence. The findings offer insights for improving maritime safety strategies and sustainable waterway transportation.

25crossref2026-05-04Gening Wang, Wenlong Zhang, Kailun Ding, Jiuteng Zhu +2 moreDOI ↗ (opens in a new tab)

Task-Coordinated Path Optimization for Grouped Unmanned Surface Vehicle Formations (opens in a new tab)

This study introduces a combined task-path optimization method for unmanned surface vehicle (USV) groups to improve efficiency by integrating task allocation and path planning. The approach uses a hierarchical framework with real-world ocean current disturbances and algorithms like an enhanced Hungarian method and SCA-APF to avoid local minima. Simulations show an 11.1% cost reduction and 80.5% better task allocation efficiency compared to genetic algorithms, with over 99% task completion in dynamic settings. The system remains fast, processing each step in under 0.1 seconds for swarms of up to 32 USVs.

26crossref2026-05-07A. Sandberg, M. Hinostroza, A. M. LekkasDOI ↗ (opens in a new tab)

Deep reinforcement learning for ship dynamic positioning: Field trials on an autonomous ferry prototype (opens in a new tab)

A deep reinforcement learning (DRL) controller using Proximal Policy Optimization (PPO) was trained and tested on a full-scale autonomous ferry prototype. The model, trained in a ROS-based 3-DOF simulator for 35 hours, was directly deployed on the physical vessel *milliAmpere1*. Field trials in Trondheim demonstrated the controller met regulatory dynamic positioning (DP) requirements, maintaining station within 4 meters and achieving ±0.25 m and ±5° accuracy after initial adjustments. Minor oscillations were attributed to velocity estimation and environmental factors.

© 2026 Ship Autonomy · Summaries are AI-generated; articles belong to their publishers · Privacy · Contact