There are several realizations of multi-agent systems of cyber-physical systems. Ocean industry and science use of heterogeneous robotic platforms is rapidly evolving. Current needs and trends fulfilling demanding marine operations objectives and tasks are stretching the robots and the corresponding sensing and control design foundation outside its comfort zone towards autonomously robot organizations and scalable hybrid methods dealing with a highly dynamic network of robots. Multi-agent systems require extreme cooperative skills, control capabilities, and resilience both as individuals and as heterogeneous robot swarms. Furthermore, multi-agent systems will enable complex missions with dynamic and adaptive planning and execution capabilities in remote and harsh environments to satisfy challenging requirements for mapping, monitoring, and intervention with unprecedented spatial and temporal resolutions and coverage. We consider this to be a major trend in all ocean operations but could indeed have applications far beyond. The talk will address various aspects and research challenges of the observation pyramid consisting of a network of sensor-carrying platforms operating in space (small-satellites), air (uncrewed aerial vehicles), sea surface (ships and autonomous surface vehicles) and underwater (autonomous and remotely controlled underwater vehicles). Experience from more than 10 years of research areas addressing topics such as design of novel concepts for marine robotics, networked operations, intelligent payload systems and sensor fusion, processing of payload data from hyperspectral imaging, artificial intelligence for decision makings using BBN and supervisory-risk control, formal and informal methods for testing and verification will be touched upon. We will show examples of conducted field campaigns. A personal reflection: We are currently approaching multi-agent systems (MAS) of autonomous cyber-physical systems (CPS) from many angles these days to investigate how to control each of them or system-of systems (network) in a decent and safe manner being able to scale up with heterogenous robotics systems operating in several different domains such as space, air, sea surface and underwater. With my own industrial and operational experience, I am concerned about both correctness by design (step 1) and secondly correctness by implementation (step 2). When we come to the latter software/hardware-in-the loop simulations will help. For MAS of autonomous CPS, I am currently more concerned about how mature the theoretical foundation is to cope with correctness by design (step 1). If we are unsure here, step 2 really doesn’t matter either. Hence, more fundamental research is needed.
Professor Asgeir J. Sørensen obtained MSc degree in Marine Technology in 1988 at NTNU, and PhD degree in Engineering Cybernetics at NTNU in 1993. In 1989-1992 Sørensen was employed at MARINTEK (SINTEF) as Research Scientist. In the years 1993-2002 Sørensen was employed in various positions in ABB whereof BA Technology Manager for Marine and Turbocharging in the global ABB Group. In December 2002 Sørensen become co-founder of the company, Marine Cybernetics AS, where he was acting as President and Chief Executive Officer (CEO) until June 2010. Marine Cybernetics AS was acquired by DNV in 2014. In 2012, 2015 and 2017 Sørensen became a co-founder of the companies Ecotone AS, Eelume AS, Pascal Technologies AS and Zeabuz AS, respectively. Since 1999 Sørensen has held the position of Professor of Marine Control Systems at the Department of Marine Technology, NTNU. In the period 2003-2012 he was key scientist in the Centre for Ships and Ocean Structures (CeSOS), and 2013-2023 Sørensen was key scientist and the Director of the Centre for Autonomous Marine Operations and Systems (NTNU AMOS). Sørensen has authored more than 280 scientific articles and book chapters, and he has graduated more than 109 MSc and 35 PhD candidates. He has also together with colleagues established the Marine Cybernetics Laboratory (MC-Lab), the Applied Underwater Robotics Laboratory (AUR-Lab) and the Observation Pyramid, all at NTNU. Sørensen is engaged in bringing fundamental research results into value creation by innovations and entrepreneurships. He is currently acting as key scientist and Director at NTNU VISTA CAROS.
