he use of digital solutions is essential to enable sustainable sea transport. To succeed with the green transition, we must leverage technological opportunities such as decision-support systems and autonomous ships to achieve energy savings and emission reductions. At the same time, increased digitalization introduces new safety challenges. The complex cause-and-effect relationships within these systems become less visible and harder to understand for both developers and operators.
Traditionally, human operators have been crucial in preventing accidents by interpreting situations and adapting to unforeseen changes. With increasing levels of digitalization and autonomy, operators’ situational awareness may weaken, reducing their ability to respond to unexpected events. SAFECOAST aims to develop methods and control architectures that can address these challenges and make autonomous ships more resilient to faults and disturbances.
A common case study has been established using DNV’s model ship Revolt, where we investigate resilient control and fault-tolerant navigation. The goal is for the ship to remain safe even when unexpected events occur. For example, it should be able to plan and operate in a way that reduces the risk of grounding even if steering or propulsion is lost. To identify potential hazardous situations, we conduct System Theoretic Process Analysis (STPA) to define safety requirements for the control systems.
A key activity is the development and testing of simulation-based safety verification methods. Instead of only monitoring the system at a high level, such as whether a collision occurs, we analyze the interactions between subsystems. This is done through contract-based design, where each subsystem is given formal assumptions and guarantees about its behavior and its interactions with the rest of the system. These contracts are tested within simulation environments, and the simulator can flag violations when assumptions are broken. This provides more precise safety analyses and can also be used as a real-time situational awareness tool onboard or in shore control centers.
The project has also analyzed autonomous ships through the lens of the High Reliability Management (HRM) framework to identify key resilience characteristics that should be built into remote operation centers and autonomous control systems. HRM emphasizes the ability to anticipate, detect, and handle unexpected events in a flexible and robust way, and this analysis provides a foundation for improving operational robustness in future autonomous shipping.
SAFECOAST officially started with a digital kick-off together with our industry partners Kongsberg Maritime and DNV, followed by the first physical project meeting in autumn 2024. All three PhD candidates are now on board, and we have established a shared platform for collaboration and method development. Together with our industry partners, we will demonstrate the developed technology through realistic case studies for autonomous ships. A new joint project meeting is planned for the end of October this year.
International collaboration is an important part of the project. We work closely with the Technical University of Denmark (DTU), where Professor Roberto Galeazzi co-supervises one of our PhD candidates. In addition, a PhD candidate from DTU is currently on an extended research stay with us, working closely with the SAFECOAST team. We also collaborate with Joanna Szlapczynska from Gdansk University of Technology in Poland on research related to resilient autonomous navigation and advanced route planning.
By combining advanced control, risk analysis, and simulation-based testing, SAFECOAST removes important barriers to digitalization and the green transition in the maritime sector. The project contributes to developing autonomous ships that are safe, reliable, and able to handle unforeseen situations — a necessary step toward a more sustainable future at sea.
Traditionally, human operators have been the main source of resilience in complex systems with their ability to interpret situations, adapt to changes, respond to disruptions to prevent accidents, and learn from experience. When computer control systems take over tasks and responsibilities from human operators, much of this resilience may be lost, since existing computer control systems are currently not able to adapt and solve problems outside their specification and design intent. The main objective of the SafeCoast project is to enable the development of highly resilient autonomous control architectures capable of coping with unstructured environments, handling sudden disturbances, deviations, failures, and making safe decisions under uncertainty. The project furthermore develops methods for understanding and assuring safe behavior of autonomous systems and their interactions with traditional systems. This is achieved by developing i) a framework for enhanced situation awareness by using online automatic hazard detection to anticipate, detect and understand system hazards; ii) a framework for online control architecture redesign that enables autonomous systems to adapt to disturbances and hazardous situations; and 3) a framework for safety assurance of adaptive and self-reconfigurable autonomous control architectures. Research results will be implemented and demonstrated in two case studies where autonomous ships are considered. An interdisciplinary approach if employed to address the research challenges from the fields of both control engineering and safety science.
