Safe and Resilient Control Systems for Autonomous Ships

The future of sustainable transportation at sea is enabled by embracing digital solutions. To ensure a successful green transition, we need to harness the power of technology, such as decision support systems and autonomous ships, to optimize energy savings and use alternative fuels. However, with increasing digitalization comes new safety challenges. The complex cause-and-effect relationships in these systems become less visible and harder to understand for both human operators and developers. Traditionally, human operators have been the backbone of resilience in complex systems, interpreting situations, adapting to changes, and preventing accidents. But as the level of system autonomy increases, human understanding and ability to provide resilience may fade, and so will the likelihood that unforeseen situations are handled. SAFECOAST is a groundbreaking research project committed to tackling these challenges. Our mission is to pave the way for the green transition by developing methods and tools that enhance safety and security at sea. Our approach involves ensuring that digital solutions not only address their intended purpose but also handle unforeseen critical situations. We aim to create resilient and fault-tolerant control architectures capable of coping with unexpected disturbances, deviations, and failures, while making safe decisions under uncertainty. The key focus is on autonomous ships, where we will develop cutting-edge technological solutions, including advanced control and decision-making systems. These systems will anticipate, detect, and understand hazardous situations, allowing for safe navigation and decision-making even amidst uncertainty. Together with industry leaders Kongsberg Maritime and DNV, who are active participants in the project, we will demonstrate this technology through case studies for autonomous ships.

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.